allstate_claims.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import lightgbm as lgb\n",
    "from datetime import datetime\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import mean_absolute_error\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn.decomposition import PCA\n",
    "from scipy import stats\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_orig = pd.read_csv('./input/train.csv')\n",
    "validate_orig = pd.read_csv('./input/test.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Based on code from here: https://www.kaggle.com/snmateen/simple-eda-feature-transformations/notebook\n",
    "\n",
    "# set max rows and columns for pandas display\n",
    "pd.set_option('display.max_columns', 1000)\n",
    "pd.set_option('display.max_rows',1000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['cat1', 'cat2', 'cat3', 'cat4', 'cat5', 'cat6', 'cat7', 'cat8', 'cat9', 'cat10', 'cat11', 'cat12', 'cat13', 'cat14', 'cat15', 'cat16', 'cat17', 'cat18', 'cat19', 'cat20', 'cat21', 'cat22', 'cat23', 'cat24', 'cat25', 'cat26', 'cat27', 'cat28', 'cat29', 'cat30', 'cat31', 'cat32', 'cat33', 'cat34', 'cat35', 'cat36', 'cat37', 'cat38', 'cat39', 'cat40', 'cat41', 'cat42', 'cat43', 'cat44', 'cat45', 'cat46', 'cat47', 'cat48', 'cat49', 'cat50', 'cat51', 'cat52', 'cat53', 'cat54', 'cat55', 'cat56', 'cat57', 'cat58', 'cat59', 'cat60', 'cat61', 'cat62', 'cat63', 'cat64', 'cat65', 'cat66', 'cat67', 'cat68', 'cat69', 'cat70', 'cat71', 'cat72', 'cat73', 'cat74', 'cat75', 'cat76', 'cat77', 'cat78', 'cat79', 'cat80', 'cat81', 'cat82', 'cat83', 'cat84', 'cat85', 'cat86', 'cat87', 'cat88', 'cat89', 'cat90', 'cat91', 'cat92', 'cat93', 'cat94', 'cat95', 'cat96', 'cat97', 'cat98', 'cat99', 'cat100', 'cat101', 'cat102', 'cat103', 'cat104', 'cat105', 'cat106', 'cat107', 'cat108', 'cat109', 'cat110', 'cat111', 'cat112', 'cat113', 'cat114', 'cat115', 'cat116']\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "116"
      ]
     },
     "execution_count": 82,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# This code adapated from here: https://stackoverflow.com/questions/22470690/get-list-of-pandas-dataframe-columns-based-on-data-type\n",
    "\n",
    "cols_obj = list(train_orig.select_dtypes(include=['object']).columns)\n",
    "print(cols_obj)\n",
    "len(cols_obj)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['id', 'cont1', 'cont2', 'cont3', 'cont4', 'cont5', 'cont6', 'cont7', 'cont8', 'cont9', 'cont10', 'cont11', 'cont12', 'cont13', 'cont14', 'loss']\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "16"
      ]
     },
     "execution_count": 83,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# This code adapated from here: https://stackoverflow.com/questions/22470690/get-list-of-pandas-dataframe-columns-based-on-data-type\n",
    "\n",
    "cols_nonobj = list(train_orig.select_dtypes(exclude=['object']).columns)\n",
    "print(cols_nonobj)\n",
    "len(cols_nonobj)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['cont1', 'cont2', 'cont3', 'cont4', 'cont5', 'cont6', 'cont7', 'cont8', 'cont9', 'cont10', 'cont11', 'cont12', 'cont13', 'cont14']\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "14"
      ]
     },
     "execution_count": 84,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cols_cont_feat = cols_nonobj[1:-1]\n",
    "print(cols_cont_feat)\n",
    "len(cols_cont_feat)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['loss']"
      ]
     },
     "execution_count": 85,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# check if any columns in one df but not the other\n",
    "np.setxor1d(train_orig.columns, validate_orig.columns).tolist()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0\n"
     ]
    }
   ],
   "source": [
    "# get count of continuous feature (non-obj columns) with missing values\n",
    "print(len(train_orig.select_dtypes(exclude=['object']).columns[train_orig.select_dtypes(exclude=['object']).isnull().sum() > 0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_loss = train_orig.loss\n",
    "# slicing columns from 1 to -1 to leave out the id and loss features: (the first and last columns of the df)\n",
    "train = train_orig.iloc[:, 1:-1]\n",
    "# slicing columns from 1 to leave out the id feature (the first column of the df)\n",
    "validate = validate_orig.iloc[:, 1:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "cat89 {'I', 'F'}\n",
      "cat90 {'G'}\n",
      "cat92 {'G', 'E', 'F'}\n",
      "cat96 {'H'}\n",
      "cat99 {'U'}\n",
      "cat101 {'U', 'N'}\n",
      "cat102 {'J', 'H'}\n",
      "cat103 {'M'}\n",
      "cat105 {'R', 'S'}\n",
      "cat106 {'Q'}\n",
      "cat109 {'BT', 'BV', 'BF', 'BM', 'BY', 'CJ', 'AG', 'AD', 'BP', 'AK', 'J', 'B'}\n",
      "cat110 {'BH', 'DV', 'H', 'BK', 'EN', 'EI', 'BN', 'BI', 'CB', 'EH', 'CA', 'AF', 'BD', 'AN'}\n",
      "cat111 {'L', 'D'}\n",
      "cat113 {'AA', 'R', 'BE', 'AC', 'T'}\n",
      "cat114 {'X'}\n",
      "cat116 {'AY', 'BF', 'DY', 'IB', 'AS', 'HS', 'MX', 'FS', 'IW', 'AB', 'W', 'HO', 'BL', 'JD', 'A', 'MB', 'GQ', 'MF', 'JT', 'IX', 'X', 'EQ', 'BN', 'DB', 'DQ', 'V', 'EX', 'FY', 'IK', 'BJ', 'KO', 'BR', 'LS', 'ER', 'BH', 'CC', 'LP', 'ET', 'HU', 'JS', 'EM', 'FO', 'AH', 'JN', 'C', 'MT', 'AT', 'AI', 'JI', 'AM', 'IS', 'EV', 'N', 'BI', 'AQ', 'AP', 'JO', 'BE', 'FN', 'IO', 'P'}\n"
     ]
    }
   ],
   "source": [
    "# Set values to NaN where a categorical value exists only in one df (train/validation)\n",
    "\n",
    "# This code adapted from here: https://www.kaggle.com/iglovikov/xgb-1114\n",
    "#    note: fixed bug in code based on use of nunique and refactored the set logic\n",
    "#    comments in code are mine\n",
    "#    prints values unique to one df col which have been set to NaN in the df in which they appear\n",
    "\n",
    "# loop thru object columns (categorical features)\n",
    "for column in list(train.select_dtypes(include=['object']).columns):\n",
    "    # get the values that exist in only one set of unique values\n",
    "    remove = set(train[column].unique()) ^ set(validate[column].unique())\n",
    "    \n",
    "    # check if the set of values to be removed is non-empty\n",
    "    if len(remove) > 0:\n",
    "        print(column, remove)\n",
    "        \n",
    "        # function to set a value to NaN in the case where it is unique to train or validate\n",
    "        def filter_cat(x):\n",
    "            if x in remove:\n",
    "                return np.nan\n",
    "            return x\n",
    "        \n",
    "        # apply the function to all values in the current column\n",
    "        train[column] = train[column].apply(filter_cat, 1)\n",
    "        validate[column] = validate[column].apply(filter_cat, 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [],
   "source": [
    "# save a reference to train before one-hot-encoding categorical columns\n",
    "train_pre_dum = train\n",
    "\n",
    "# one-hot-encode categorical features of train df (get_dummies ignores the floating point columns)\n",
    "train = pd.get_dummies(train)\n",
    "\n",
    "# one-hot-encode categorical features of validation df (get_dummies ignores the floating point columns)\n",
    "validate = pd.get_dummies(validate)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "set()"
      ]
     },
     "execution_count": 90,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# check if any columns in one df but not the other\n",
    "#np.setxor1d(train.columns, validate.columns).tolist()\n",
    "set(train.columns) ^ set(validate.columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [],
   "source": [
    "# create a lightGBM regressor object\n",
    "lgbm_reg = lgb.LGBMRegressor(random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [],
   "source": [
    "# split data into train and test dfs\n",
    "X_train, X_test, y_train, y_test = train_test_split(train, train_loss, test_size=.2, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0:00:07.856895\n",
      "The mae of the model prediction is:  1601.87628096\n"
     ]
    }
   ],
   "source": [
    "# Fit the model using default parameters and predict the target (loss) values based on the model\n",
    "#  then compare prediction with actual values and compute mean absolute error of residuals\n",
    "\n",
    "# This is based on code from https://www.analyticsvidhya.com/blog/2017/06/which-algorithm-takes-the-crown-light-gbm-vs-xgboost\n",
    "\n",
    "start=datetime.now()\n",
    "\n",
    "lgbm_reg.fit(X_train, y_train)\n",
    "        \n",
    "stop=datetime.now()\n",
    "print(stop-start)\n",
    "\n",
    "y_pred=lgbm_reg.predict(X_test)\n",
    "print('The mae of the model prediction is: ', mean_absolute_error(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0:00:47.985074\n",
      "The mae of the model prediction is:  1211.80536662\n",
      "\n",
      "Best params:\n",
      "{'learning_rate': 1.2, 'min_data_in_leaf': 500, 'n_estimators': 110, 'num_leaves': 50}\n"
     ]
    }
   ],
   "source": [
    "# Final parameters\n",
    "\n",
    "param_grid = {\n",
    "    'learning_rate': [1.2],\n",
    "    'n_estimators': [110],\n",
    "    'num_leaves': [50],\n",
    "    'min_data_in_leaf': [500]\n",
    "}\n",
    "\n",
    "start=datetime.now()\n",
    "\n",
    "lgbm_reg_best = GridSearchCV(lgbm_reg, param_grid)\n",
    "lgbm_reg_best.fit(X_train, y_train)\n",
    "\n",
    "stop=datetime.now()\n",
    "print(stop-start)\n",
    "\n",
    "y_pred=lgbm_reg_best.predict(X_test)\n",
    "print('The mae of the model prediction is: ', mean_absolute_error(y_test, y_pred))\n",
    "\n",
    "print('\\nBest params:\\n{}'.format(lgbm_reg_best.best_params_))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Anaconda3\\lib\\site-packages\\seaborn\\categorical.py:598: FutureWarning: remove_na is deprecated and is a private function. Do not use.\n",
      "  kde_data = remove_na(group_data)\n",
      "C:\\Anaconda3\\lib\\site-packages\\seaborn\\categorical.py:826: FutureWarning: remove_na is deprecated and is a private function. Do not use.\n",
      "  violin_data = remove_na(group_data)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZsAAADuCAYAAADvAKo+AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n/wPBLeXFVcpK9AAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x27304c12780>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Graph showing skew of loss data\n",
    "\n",
    "# This code is adapted from here: https://www.kaggle.com/sharmasanthosh/exploratory-study-on-ml-algorithms/notebook\n",
    "\n",
    "sns.violinplot(data=train_orig, y=\"loss\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Adjust the highly-skewed loss data\n",
    "\n",
    "#This code is from here: https://www.kaggle.com/iglovikov/xgb-1114\n",
    "#  The shift value was discovered empirically\n",
    "shift = 200\n",
    "train_loss = np.log(train_orig[\"loss\"] + shift)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Anaconda3\\lib\\site-packages\\seaborn\\categorical.py:598: FutureWarning: remove_na is deprecated and is a private function. Do not use.\n",
      "  kde_data = remove_na(group_data)\n",
      "C:\\Anaconda3\\lib\\site-packages\\seaborn\\categorical.py:826: FutureWarning: remove_na is deprecated and is a private function. Do not use.\n",
      "  violin_data = remove_na(group_data)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAAD8CAYAAABn919SAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nEBLoQggRIf4/Hjx7vC3edwgAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x27305898dd8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# graph the data after transform\n",
    "sns.violinplot(data=train_loss)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {},
   "outputs": [],
   "source": [
    "# split data into train and test dfs\n",
    "X_train, X_test, y_train, y_test = train_test_split(train, train_loss, test_size=.2, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0:00:51.085867\n",
      "The mae of the model prediction is:  1146.1330493\n",
      "\n",
      "Best params:\n",
      "{'learning_rate': 0.2, 'min_data_in_leaf': 300, 'n_estimators': 300, 'num_leaves': 10}\n"
     ]
    }
   ],
   "source": [
    "# Final parameters\n",
    "\n",
    "param_grid = {\n",
    "    'learning_rate': [.2],\n",
    "    'n_estimators': [300],\n",
    "    'num_leaves': [10],\n",
    "    'min_data_in_leaf': [300]\n",
    "}\n",
    "\n",
    "start=datetime.now()\n",
    "\n",
    "lgbm_reg_best = GridSearchCV(lgbm_reg, param_grid)\n",
    "lgbm_reg_best.fit(X_train, y_train)\n",
    "\n",
    "stop=datetime.now()\n",
    "print(stop-start)\n",
    "\n",
    "y_pred=lgbm_reg_best.predict(X_test)\n",
    "\n",
    "# adjust the predictions (due to adjustment of loss)\n",
    "adj_y_test = np.exp(y_test) - shift\n",
    "adj_y_pred = np.exp(y_pred) - shift\n",
    "\n",
    "#print('The mae of the model prediction is: ', mean_absolute_error(y_test, y_pred))\n",
    "print('The mae of the model prediction is: ', mean_absolute_error(adj_y_test, adj_y_pred))\n",
    "\n",
    "print('\\nBest params:\\n{}'.format(lgbm_reg_best.best_params_))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "cont11  cont12    0.994384\n",
       "cont1   cont9     0.929912\n",
       "cont10  cont6     0.883351\n",
       "cont13  cont6     0.815091\n",
       "cont1   cont10    0.808551\n",
       "cont6   cont9     0.797544\n",
       "cont10  cont9     0.785697\n",
       "cont12  cont6     0.785144\n",
       "cont11  cont6     0.773745\n",
       "cont1   cont6     0.758315\n",
       "dtype: float64"
      ]
     },
     "execution_count": 100,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Remove highly correlated features\n",
    "\n",
    "# Highly adapted from code here: https://www.kaggle.com/sharmasanthosh/exploratory-study-on-ml-algorithms/notebook\n",
    "# Calculates pearson co-efficient for all combinations\n",
    "train_corr = train_orig.iloc[:,1:-1].corr(method='pearson').stack()\n",
    "# This was a helpful example of selecting every other row in df: https://stackoverflow.com/questions/37280681/reverse-even-rows-in-a-numpy-array-or-pandas-dataframe\n",
    "train_corr[(train_corr < 1) & (train_corr >= .75)].sort_values(ascending=False).iloc[1::2]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['cont3', 'cont7', 'cont4', 'cont2', 'cont5', 'cont14', 'cont8', 'cont13', 'cont6']\n"
     ]
    }
   ],
   "source": [
    "# Since: \n",
    "#     cont6 --> cont10, 13, 9, and 12\n",
    "#     and cont9 & 10 --> cont1\n",
    "#     and cont12 --> cont11\n",
    "# Therefore:\n",
    "#     Continuous features 9, 10, 11 & 12 are redundant with cont6\n",
    "\n",
    "cols_cont_drop = ['cont1','cont9','cont10','cont11','cont12']\n",
    "\n",
    "# This example was helpful here: https://stackoverflow.com/questions/13003051/how-do-i-exclude-a-few-columns-from-a-dataframe-plot\n",
    "cols_remnant = train.columns.difference(cols_cont_drop)\n",
    "\n",
    "cols_cont_feat = list(set(cols_cont_feat) - set(cols_cont_drop))\n",
    "\n",
    "print(cols_cont_feat)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>cont3</th>\n",
       "      <th>cont7</th>\n",
       "      <th>cont4</th>\n",
       "      <th>cont2</th>\n",
       "      <th>cont5</th>\n",
       "      <th>cont14</th>\n",
       "      <th>cont8</th>\n",
       "      <th>cont13</th>\n",
       "      <th>cont6</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>188318.000000</td>\n",
       "      <td>188318.000000</td>\n",
       "      <td>188318.000000</td>\n",
       "      <td>188318.000000</td>\n",
       "      <td>188318.000000</td>\n",
       "      <td>188318.000000</td>\n",
       "      <td>188318.000000</td>\n",
       "      <td>188318.000000</td>\n",
       "      <td>188318.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>0.498918</td>\n",
       "      <td>0.484970</td>\n",
       "      <td>0.491812</td>\n",
       "      <td>0.507188</td>\n",
       "      <td>0.487428</td>\n",
       "      <td>0.495717</td>\n",
       "      <td>0.486437</td>\n",
       "      <td>0.493138</td>\n",
       "      <td>0.490945</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.202105</td>\n",
       "      <td>0.178450</td>\n",
       "      <td>0.211292</td>\n",
       "      <td>0.207202</td>\n",
       "      <td>0.209027</td>\n",
       "      <td>0.222488</td>\n",
       "      <td>0.199370</td>\n",
       "      <td>0.212777</td>\n",
       "      <td>0.205273</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.002634</td>\n",
       "      <td>0.069503</td>\n",
       "      <td>0.176921</td>\n",
       "      <td>0.001149</td>\n",
       "      <td>0.281143</td>\n",
       "      <td>0.179722</td>\n",
       "      <td>0.236880</td>\n",
       "      <td>0.000228</td>\n",
       "      <td>0.012683</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>0.336963</td>\n",
       "      <td>0.350175</td>\n",
       "      <td>0.327354</td>\n",
       "      <td>0.358319</td>\n",
       "      <td>0.281143</td>\n",
       "      <td>0.294610</td>\n",
       "      <td>0.312800</td>\n",
       "      <td>0.315758</td>\n",
       "      <td>0.336105</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>0.527991</td>\n",
       "      <td>0.438285</td>\n",
       "      <td>0.452887</td>\n",
       "      <td>0.555782</td>\n",
       "      <td>0.422268</td>\n",
       "      <td>0.407403</td>\n",
       "      <td>0.441060</td>\n",
       "      <td>0.363547</td>\n",
       "      <td>0.440945</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>0.634224</td>\n",
       "      <td>0.591045</td>\n",
       "      <td>0.652072</td>\n",
       "      <td>0.681761</td>\n",
       "      <td>0.643315</td>\n",
       "      <td>0.724623</td>\n",
       "      <td>0.623580</td>\n",
       "      <td>0.689974</td>\n",
       "      <td>0.655021</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>0.944251</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.954297</td>\n",
       "      <td>0.862654</td>\n",
       "      <td>0.983674</td>\n",
       "      <td>0.844848</td>\n",
       "      <td>0.980200</td>\n",
       "      <td>0.988494</td>\n",
       "      <td>0.997162</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               cont3          cont7          cont4          cont2  \\\n",
       "count  188318.000000  188318.000000  188318.000000  188318.000000   \n",
       "mean        0.498918       0.484970       0.491812       0.507188   \n",
       "std         0.202105       0.178450       0.211292       0.207202   \n",
       "min         0.002634       0.069503       0.176921       0.001149   \n",
       "25%         0.336963       0.350175       0.327354       0.358319   \n",
       "50%         0.527991       0.438285       0.452887       0.555782   \n",
       "75%         0.634224       0.591045       0.652072       0.681761   \n",
       "max         0.944251       1.000000       0.954297       0.862654   \n",
       "\n",
       "               cont5         cont14          cont8         cont13  \\\n",
       "count  188318.000000  188318.000000  188318.000000  188318.000000   \n",
       "mean        0.487428       0.495717       0.486437       0.493138   \n",
       "std         0.209027       0.222488       0.199370       0.212777   \n",
       "min         0.281143       0.179722       0.236880       0.000228   \n",
       "25%         0.281143       0.294610       0.312800       0.315758   \n",
       "50%         0.422268       0.407403       0.441060       0.363547   \n",
       "75%         0.643315       0.724623       0.623580       0.689974   \n",
       "max         0.983674       0.844848       0.980200       0.988494   \n",
       "\n",
       "               cont6  \n",
       "count  188318.000000  \n",
       "mean        0.490945  \n",
       "std         0.205273  \n",
       "min         0.012683  \n",
       "25%         0.336105  \n",
       "50%         0.440945  \n",
       "75%         0.655021  \n",
       "max         0.997162  "
      ]
     },
     "execution_count": 102,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train = train[cols_remnant]\n",
    "train[cols_cont_feat].describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {},
   "outputs": [],
   "source": [
    "# split data into train and test dfs\n",
    "X_train, X_test, y_train, y_test = train_test_split(train, train_loss, test_size=.2, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0:01:14.533219\n",
      "The mae of the model prediction is:  1139.19598328\n",
      "\n",
      "Best params:\n",
      "{'learning_rate': 0.1, 'min_data_in_leaf': 125, 'n_estimators': 500, 'num_leaves': 16}\n"
     ]
    }
   ],
   "source": [
    "# Final parameters\n",
    "\n",
    "param_grid = {\n",
    "    'learning_rate': [.1],\n",
    "    'n_estimators': [500],\n",
    "    'num_leaves': [16],\n",
    "    'min_data_in_leaf': [125]\n",
    "}\n",
    "\n",
    "start=datetime.now()\n",
    "\n",
    "lgbm_reg_best = GridSearchCV(lgbm_reg, param_grid)\n",
    "lgbm_reg_best.fit(X_train, y_train)\n",
    "\n",
    "stop=datetime.now()\n",
    "print(stop-start)\n",
    "\n",
    "y_pred=lgbm_reg_best.predict(X_test)\n",
    "\n",
    "# adjust the predictions (due to adjustment of loss)\n",
    "adj_y_test = np.exp(y_test) - shift\n",
    "adj_y_pred = np.exp(y_pred) - shift\n",
    "\n",
    "#print('The mae of the model prediction is: ', mean_absolute_error(y_test, y_pred))\n",
    "print('The mae of the model prediction is: ', mean_absolute_error(adj_y_test, adj_y_pred))\n",
    "\n",
    "print('\\nBest params:\\n{}'.format(lgbm_reg_best.best_params_))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>cont3</th>\n",
       "      <th>cont7</th>\n",
       "      <th>cont4</th>\n",
       "      <th>cont2</th>\n",
       "      <th>cont5</th>\n",
       "      <th>cont14</th>\n",
       "      <th>cont8</th>\n",
       "      <th>cont13</th>\n",
       "      <th>cont6</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>188318</td>\n",
       "      <td>188318</td>\n",
       "      <td>188318</td>\n",
       "      <td>188318</td>\n",
       "      <td>188318</td>\n",
       "      <td>188318</td>\n",
       "      <td>188318</td>\n",
       "      <td>188318</td>\n",
       "      <td>188318</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>unique</th>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>top</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>freq</th>\n",
       "      <td>188318</td>\n",
       "      <td>185659</td>\n",
       "      <td>188318</td>\n",
       "      <td>188318</td>\n",
       "      <td>188318</td>\n",
       "      <td>188318</td>\n",
       "      <td>188318</td>\n",
       "      <td>188318</td>\n",
       "      <td>188318</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         cont3   cont7   cont4   cont2   cont5  cont14   cont8  cont13   cont6\n",
       "count   188318  188318  188318  188318  188318  188318  188318  188318  188318\n",
       "unique       1       2       1       1       1       1       1       1       1\n",
       "top      False   False   False   False   False   False   False   False   False\n",
       "freq    188318  185659  188318  188318  188318  188318  188318  188318  188318"
      ]
     },
     "execution_count": 105,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Use 1.5 * IQR to define outliners\n",
    "\n",
    "def get_outliers(series):\n",
    "    Q1 = np.percentile(series, 25)\n",
    "    Q3 = np.percentile(series, 75)\n",
    "    IQR = Q3-Q1\n",
    "    step = IQR * 1.5\n",
    "    lower_bound = Q1 - step\n",
    "    upper_bound = Q3 + step\n",
    "    \n",
    "    return series.apply(lambda x: not lower_bound <= x <= upper_bound)\n",
    "\n",
    "outliers = pd.DataFrame()\n",
    "\n",
    "for col in cols_cont_feat:\n",
    "    col_outs = get_outliers(train[col])\n",
    "    outliers = pd.concat([outliers, col_outs], axis=1)\n",
    "\n",
    "total_count, _ = outliers.shape\n",
    "outliers.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2659\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.014119733642031033"
      ]
     },
     "execution_count": 106,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# look at count of outliers and pct of total values\n",
    "\n",
    "out_count = outliers.cont7.value_counts()[1]\n",
    "print(out_count)\n",
    "out_count / total_count"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.43828500000000004\n",
      "2659\n",
      "77 0.43828500000000004 0.985762\n",
      "106 0.43828500000000004 0.982302\n",
      "179 0.43828500000000004 0.993938\n",
      "275 0.43828500000000004 0.95483\n",
      "301 0.43828500000000004 0.997761\n",
      "327 0.43828500000000004 0.98473\n",
      "369 0.43828500000000004 0.953884\n",
      "406 0.43828500000000004 0.965359\n",
      "490 0.43828500000000004 0.999259\n",
      "503 0.43828500000000004 0.979276\n",
      "520 0.43828500000000004 0.970764\n",
      "541 0.43828500000000004 0.989659\n",
      "731 0.43828500000000004 0.95483\n",
      "791 0.43828500000000004 0.99792\n",
      "879 0.43828500000000004 0.972119\n",
      "922 0.43828500000000004 0.999708\n",
      "924 0.43828500000000004 0.952897\n",
      "989 0.43828500000000004 0.986583\n",
      "1031 0.43828500000000004 0.978156\n",
      "1053 0.43828500000000004 0.982302\n",
      "1144 0.43828500000000004 0.997098\n",
      "1301 0.43828500000000004 0.99884\n",
      "1389 0.43828500000000004 0.966972\n",
      "1536 0.43828500000000004 0.964994\n",
      "1543 0.43828500000000004 0.972119\n",
      "1624 0.43828500000000004 0.960335\n",
      "1632 0.43828500000000004 0.977482\n",
      "1719 0.43828500000000004 0.99433\n",
      "1737 0.43828500000000004 0.997734\n",
      "1776 0.43828500000000004 0.997299\n",
      "1819 0.43828500000000004 0.996914\n",
      "1836 0.43828500000000004 1.0\n",
      "1924 0.43828500000000004 0.989003\n",
      "1941 0.43828500000000004 0.972119\n",
      "1943 0.43828500000000004 0.976141\n",
      "1997 0.43828500000000004 0.970764\n",
      "2082 0.43828500000000004 0.965011\n",
      "2139 0.43828500000000004 0.995395\n",
      "2161 0.43828500000000004 0.997306\n",
      "2270 0.43828500000000004 0.984877\n",
      "2322 0.43828500000000004 0.980178\n",
      "2366 0.43828500000000004 0.969344\n",
      "2401 0.43828500000000004 0.95519\n",
      "2572 0.43828500000000004 0.977514\n",
      "2582 0.43828500000000004 0.954936\n",
      "2672 0.43828500000000004 0.999021\n",
      "2718 0.43828500000000004 0.989659\n",
      "2770 0.43828500000000004 0.971222\n",
      "2781 0.43828500000000004 0.974758\n",
      "2835 0.43828500000000004 0.990853\n",
      "2886 0.43828500000000004 0.963166\n",
      "2972 0.43828500000000004 0.982293\n",
      "2977 0.43828500000000004 0.953162\n",
      "3054 0.43828500000000004 0.995843\n",
      "3093 0.43828500000000004 0.996931\n",
      "3106 0.43828500000000004 0.975653\n",
      "3225 0.43828500000000004 0.999998\n",
      "3247 0.43828500000000004 0.95483\n",
      "3279 0.43828500000000004 0.99837\n",
      "3370 0.43828500000000004 0.960071\n",
      "3446 0.43828500000000004 0.999021\n",
      "3466 0.43828500000000004 0.955799\n",
      "3478 0.43828500000000004 0.960071\n",
      "3590 0.43828500000000004 0.985969\n",
      "3619 0.43828500000000004 0.972119\n",
      "3715 0.43828500000000004 0.99805\n",
      "3795 0.43828500000000004 0.999551\n",
      "3851 0.43828500000000004 0.981646\n",
      "3920 0.43828500000000004 0.96192\n",
      "4013 0.43828500000000004 0.998292\n",
      "4174 0.43828500000000004 0.990343\n",
      "4397 0.43828500000000004 0.961043\n",
      "4477 0.43828500000000004 0.970917\n",
      "4503 0.43828500000000004 0.968527\n",
      "4632 0.43828500000000004 0.990343\n",
      "4727 0.43828500000000004 0.986583\n",
      "4803 0.43828500000000004 0.988151\n",
      "5005 0.43828500000000004 0.994097\n",
      "5086 0.43828500000000004 0.956727\n",
      "5102 0.43828500000000004 0.997553\n",
      "5212 0.43828500000000004 0.971944\n",
      "5219 0.43828500000000004 0.986803\n",
      "5253 0.43828500000000004 0.997098\n",
      "5258 0.43828500000000004 0.987662\n",
      "5269 0.43828500000000004 0.980111\n",
      "5316 0.43828500000000004 0.973489\n",
      "5341 0.43828500000000004 0.969344\n",
      "5409 0.43828500000000004 0.985574\n",
      "5458 0.43828500000000004 0.992198\n",
      "5548 0.43828500000000004 0.999891\n",
      "5559 0.43828500000000004 0.994888\n",
      "5578 0.43828500000000004 0.978156\n",
      "5681 0.43828500000000004 0.964202\n",
      "5717 0.43828500000000004 0.997909\n",
      "5729 0.43828500000000004 0.987033\n",
      "5894 0.43828500000000004 0.957033\n",
      "5941 0.43828500000000004 0.970369\n",
      "6154 0.43828500000000004 0.994097\n",
      "6163 0.43828500000000004 0.986386\n",
      "6191 0.43828500000000004 0.970369\n",
      "6320 0.43828500000000004 0.980359\n",
      "6346 0.43828500000000004 0.986803\n",
      "6462 0.43828500000000004 0.966972\n",
      "6513 0.43828500000000004 0.95519\n",
      "6536 0.43828500000000004 0.994229\n",
      "6636 0.43828500000000004 0.990343\n",
      "6666 0.43828500000000004 0.982302\n",
      "6676 0.43828500000000004 0.970764\n",
      "6677 0.43828500000000004 0.998283\n",
      "6690 0.43828500000000004 0.989067\n",
      "6702 0.43828500000000004 0.987799\n",
      "6733 0.43828500000000004 1.0\n",
      "6825 0.43828500000000004 0.997909\n",
      "6952 0.43828500000000004 0.988543\n",
      "6973 0.43828500000000004 0.991667\n",
      "7064 0.43828500000000004 0.997479\n",
      "7332 0.43828500000000004 0.960071\n",
      "7380 0.43828500000000004 0.996613\n",
      "7516 0.43828500000000004 0.970749\n",
      "7518 0.43828500000000004 0.982302\n",
      "7628 0.43828500000000004 0.977482\n",
      "7666 0.43828500000000004 0.994229\n",
      "7760 0.43828500000000004 0.987583\n",
      "7785 0.43828500000000004 0.978563\n",
      "7839 0.43828500000000004 0.998678\n",
      "7879 0.43828500000000004 0.998525\n",
      "8145 0.43828500000000004 0.980539\n",
      "8160 0.43828500000000004 0.994212\n",
      "8220 0.43828500000000004 0.983131\n",
      "8257 0.43828500000000004 0.97773\n",
      "8316 0.43828500000000004 0.970141\n",
      "8360 0.43828500000000004 0.994097\n",
      "8500 0.43828500000000004 0.990343\n",
      "8524 0.43828500000000004 1.0\n",
      "8552 0.43828500000000004 0.953162\n",
      "8606 0.43828500000000004 0.972119\n",
      "8743 0.43828500000000004 0.994883\n",
      "8768 0.43828500000000004 0.989659\n",
      "8823 0.43828500000000004 0.970917\n",
      "8862 0.43828500000000004 0.960541\n",
      "8868 0.43828500000000004 0.966909\n",
      "8879 0.43828500000000004 0.97773\n",
      "8896 0.43828500000000004 0.952897\n",
      "9503 0.43828500000000004 0.964994\n",
      "9505 0.43828500000000004 0.997704\n",
      "9537 0.43828500000000004 0.998444\n",
      "9687 0.43828500000000004 0.981753\n",
      "9825 0.43828500000000004 0.994883\n",
      "9876 0.43828500000000004 0.996613\n",
      "9891 0.43828500000000004 0.993983\n",
      "9920 0.43828500000000004 0.963687\n",
      "9936 0.43828500000000004 0.984877\n",
      "10070 0.43828500000000004 0.970764\n",
      "10075 0.43828500000000004 0.997999\n",
      "10329 0.43828500000000004 0.997429\n",
      "10339 0.43828500000000004 0.970749\n",
      "10492 0.43828500000000004 0.997339\n",
      "10516 0.43828500000000004 0.979585\n",
      "10645 0.43828500000000004 0.983962\n",
      "10661 0.43828500000000004 0.998684\n",
      "10684 0.43828500000000004 0.999999\n",
      "10749 0.43828500000000004 0.985574\n",
      "10777 0.43828500000000004 0.973705\n",
      "10844 0.43828500000000004 0.997999\n",
      "10941 0.43828500000000004 0.994421\n",
      "11062 0.43828500000000004 0.982302\n",
      "11118 0.43828500000000004 0.994994\n",
      "11265 0.43828500000000004 0.973566\n",
      "11281 0.43828500000000004 0.997734\n",
      "11341 0.43828500000000004 0.960541\n",
      "11375 0.43828500000000004 0.970749\n",
      "11443 0.43828500000000004 0.960429\n",
      "11534 0.43828500000000004 0.987184\n",
      "11605 0.43828500000000004 0.962726\n",
      "11616 0.43828500000000004 0.998537\n",
      "11619 0.43828500000000004 0.96192\n",
      "11651 0.43828500000000004 0.99837\n",
      "11671 0.43828500000000004 0.996931\n",
      "11912 0.43828500000000004 0.994545\n",
      "11939 0.43828500000000004 0.985205\n",
      "11958 0.43828500000000004 0.964541\n",
      "12014 0.43828500000000004 0.952897\n",
      "12137 0.43828500000000004 0.993009\n",
      "12399 0.43828500000000004 0.964541\n",
      "12524 0.43828500000000004 0.991681\n",
      "12552 0.43828500000000004 0.972596\n",
      "12605 0.43828500000000004 0.960541\n",
      "12619 0.43828500000000004 0.997761\n",
      "12708 0.43828500000000004 0.987516\n",
      "12856 0.43828500000000004 0.959423\n",
      "12863 0.43828500000000004 0.955861\n",
      "12995 0.43828500000000004 0.994369\n",
      "13030 0.43828500000000004 1.0\n",
      "13064 0.43828500000000004 0.983679\n",
      "13194 0.43828500000000004 0.993063\n",
      "13297 0.43828500000000004 0.97162\n",
      "13340 0.43828500000000004 0.974379\n",
      "13463 0.43828500000000004 0.975465\n",
      "13548 0.43828500000000004 0.973001\n",
      "13711 0.43828500000000004 0.998444\n",
      "13878 0.43828500000000004 0.965359\n",
      "14172 0.43828500000000004 0.985762\n",
      "14357 0.43828500000000004 0.987799\n",
      "14384 0.43828500000000004 0.988841\n",
      "14385 0.43828500000000004 0.970749\n",
      "14404 0.43828500000000004 0.999573\n",
      "14472 0.43828500000000004 0.964338\n",
      "14689 0.43828500000000004 0.95519\n",
      "14811 0.43828500000000004 0.973566\n",
      "14879 0.43828500000000004 0.988928\n",
      "14913 0.43828500000000004 0.997999\n",
      "14969 0.43828500000000004 1.0\n",
      "15016 0.43828500000000004 0.993983\n",
      "15041 0.43828500000000004 0.990343\n",
      "15049 0.43828500000000004 0.99805\n",
      "15054 0.43828500000000004 0.999095\n",
      "15152 0.43828500000000004 0.970693\n",
      "15215 0.43828500000000004 0.993938\n",
      "15293 0.43828500000000004 0.992531\n",
      "15323 0.43828500000000004 0.968467\n",
      "15346 0.43828500000000004 0.972936\n",
      "15425 0.43828500000000004 0.953775\n",
      "15431 0.43828500000000004 0.984377\n",
      "15438 0.43828500000000004 0.954936\n",
      "15513 0.43828500000000004 0.98969\n",
      "15630 0.43828500000000004 0.989424\n",
      "15823 0.43828500000000004 0.982472\n",
      "15887 0.43828500000000004 0.991396\n",
      "15930 0.43828500000000004 0.985969\n",
      "15995 0.43828500000000004 0.983131\n",
      "16066 0.43828500000000004 0.979761\n",
      "16105 0.43828500000000004 0.992198\n",
      "16246 0.43828500000000004 1.0\n",
      "16507 0.43828500000000004 0.998257\n",
      "16543 0.43828500000000004 0.960354\n",
      "16550 0.43828500000000004 0.999936\n",
      "16616 0.43828500000000004 0.997734\n",
      "16717 0.43828500000000004 0.987516\n",
      "17186 0.43828500000000004 0.99857\n",
      "17430 0.43828500000000004 0.990343\n",
      "17464 0.43828500000000004 0.988543\n",
      "17586 0.43828500000000004 0.985735\n",
      "17618 0.43828500000000004 0.994845\n",
      "17691 0.43828500000000004 0.990343\n",
      "17693 0.43828500000000004 0.960071\n",
      "17715 0.43828500000000004 0.974514\n",
      "17987 0.43828500000000004 0.991667\n",
      "18048 0.43828500000000004 0.991866\n",
      "18058 0.43828500000000004 0.993009\n",
      "18271 0.43828500000000004 0.982302\n",
      "18342 0.43828500000000004 0.980642\n",
      "18367 0.43828500000000004 0.964304\n",
      "18457 0.43828500000000004 0.996914\n",
      "18645 0.43828500000000004 0.997429\n",
      "18682 0.43828500000000004 0.998382\n",
      "18821 0.43828500000000004 0.972119\n",
      "18905 0.43828500000000004 0.976652\n",
      "18955 0.43828500000000004 0.965359\n",
      "19037 0.43828500000000004 0.973566\n",
      "19083 0.43828500000000004 0.997734\n",
      "19090 0.43828500000000004 0.998444\n",
      "19299 0.43828500000000004 0.983016\n",
      "19324 0.43828500000000004 0.999147\n",
      "19349 0.43828500000000004 0.970764\n",
      "19399 0.43828500000000004 0.960071\n",
      "19492 0.43828500000000004 0.977482\n",
      "19498 0.43828500000000004 0.983464\n",
      "19533 0.43828500000000004 0.981753\n",
      "19674 0.43828500000000004 0.972557\n",
      "19699 0.43828500000000004 0.952897\n",
      "19768 0.43828500000000004 0.972557\n",
      "19835 0.43828500000000004 0.973566\n",
      "19859 0.43828500000000004 0.993938\n",
      "19917 0.43828500000000004 0.985735\n",
      "20004 0.43828500000000004 0.973001\n",
      "20142 0.43828500000000004 0.970749\n",
      "20238 0.43828500000000004 0.967767\n",
      "20263 0.43828500000000004 0.998444\n",
      "20275 0.43828500000000004 0.998463\n",
      "20280 0.43828500000000004 0.993009\n",
      "20399 0.43828500000000004 0.997734\n",
      "20460 0.43828500000000004 0.972557\n",
      "20725 0.43828500000000004 0.990853\n",
      "20813 0.43828500000000004 0.972119\n",
      "20896 0.43828500000000004 0.997999\n",
      "20928 0.43828500000000004 0.992446\n",
      "20956 0.43828500000000004 0.976175\n",
      "21099 0.43828500000000004 0.957033\n",
      "21180 0.43828500000000004 0.966972\n",
      "21315 0.43828500000000004 0.992198\n",
      "21348 0.43828500000000004 0.975944\n",
      "21399 0.43828500000000004 0.962637\n",
      "21493 0.43828500000000004 0.99462\n",
      "21581 0.43828500000000004 0.987811\n",
      "21652 0.43828500000000004 0.981806\n",
      "21757 0.43828500000000004 0.955169\n",
      "21760 0.43828500000000004 0.993724\n",
      "21764 0.43828500000000004 0.987553\n",
      "21784 0.43828500000000004 0.96192\n",
      "21822 0.43828500000000004 0.970369\n",
      "21871 0.43828500000000004 0.989409\n",
      "21922 0.43828500000000004 0.988029\n",
      "21983 0.43828500000000004 0.986681\n",
      "21988 0.43828500000000004 0.981023\n",
      "22026 0.43828500000000004 0.980539\n",
      "22219 0.43828500000000004 0.999984\n",
      "22287 0.43828500000000004 0.967767\n",
      "22360 0.43828500000000004 0.998444\n",
      "22459 0.43828500000000004 0.987739\n",
      "22476 0.43828500000000004 0.959423\n",
      "22511 0.43828500000000004 0.98944\n",
      "22637 0.43828500000000004 0.95519\n",
      "22645 0.43828500000000004 0.970749\n",
      "22686 0.43828500000000004 0.987184\n",
      "23099 0.43828500000000004 0.993063\n",
      "23204 0.43828500000000004 0.988543\n",
      "23448 0.43828500000000004 0.95851\n",
      "23480 0.43828500000000004 0.960071\n",
      "23698 0.43828500000000004 0.985762\n",
      "23735 0.43828500000000004 0.999953\n",
      "23754 0.43828500000000004 0.994545\n",
      "23862 0.43828500000000004 0.967767\n",
      "23869 0.43828500000000004 0.999021\n",
      "23990 0.43828500000000004 0.997693\n",
      "24188 0.43828500000000004 0.954936\n",
      "24399 0.43828500000000004 0.992229\n",
      "24712 0.43828500000000004 0.957937\n",
      "24760 0.43828500000000004 0.957937\n",
      "24764 0.43828500000000004 0.960541\n",
      "24782 0.43828500000000004 0.970369\n",
      "24828 0.43828500000000004 0.979585\n",
      "24834 0.43828500000000004 0.993724\n",
      "24850 0.43828500000000004 0.999095\n",
      "24859 0.43828500000000004 0.99858\n",
      "24982 0.43828500000000004 0.99607\n",
      "25049 0.43828500000000004 0.991706\n",
      "25094 0.43828500000000004 0.997339\n",
      "25292 0.43828500000000004 0.993508\n",
      "25328 0.43828500000000004 0.99871\n",
      "25479 0.43828500000000004 0.999842\n",
      "25598 0.43828500000000004 0.95519\n",
      "25602 0.43828500000000004 0.992198\n",
      "25616 0.43828500000000004 0.989948\n",
      "25718 0.43828500000000004 1.0\n",
      "25730 0.43828500000000004 0.987184\n",
      "25800 0.43828500000000004 0.968073\n",
      "26130 0.43828500000000004 0.952897\n",
      "26138 0.43828500000000004 0.992289\n",
      "26154 0.43828500000000004 0.957033\n",
      "26170 0.43828500000000004 0.989424\n",
      "26191 0.43828500000000004 0.970749\n",
      "26196 0.43828500000000004 0.952963\n",
      "26246 0.43828500000000004 0.957033\n",
      "26266 0.43828500000000004 0.953884\n",
      "26342 0.43828500000000004 0.96192\n",
      "26443 0.43828500000000004 0.999817\n",
      "26641 0.43828500000000004 0.953884\n",
      "26663 0.43828500000000004 0.984811\n",
      "26714 0.43828500000000004 0.984877\n",
      "26850 0.43828500000000004 0.968467\n",
      "26859 0.43828500000000004 0.985762\n",
      "26894 0.43828500000000004 0.95584\n",
      "26937 0.43828500000000004 0.965011\n",
      "26984 0.43828500000000004 0.969344\n",
      "27002 0.43828500000000004 0.976675\n",
      "27037 0.43828500000000004 0.98386\n",
      "27181 0.43828500000000004 0.972119\n",
      "27271 0.43828500000000004 0.988928\n",
      "27285 0.43828500000000004 0.985454\n",
      "27326 0.43828500000000004 0.982302\n",
      "27524 0.43828500000000004 0.977482\n",
      "27545 0.43828500000000004 0.993983\n",
      "27692 0.43828500000000004 0.990853\n",
      "28009 0.43828500000000004 0.999115\n",
      "28046 0.43828500000000004 0.994421\n",
      "28138 0.43828500000000004 0.987033\n",
      "28274 0.43828500000000004 0.979693\n",
      "28285 0.43828500000000004 0.973566\n",
      "28324 0.43828500000000004 0.952897\n",
      "28374 0.43828500000000004 0.961283\n",
      "28434 0.43828500000000004 0.973001\n",
      "28535 0.43828500000000004 0.978156\n",
      "28547 0.43828500000000004 0.987553\n",
      "28561 0.43828500000000004 0.983464\n",
      "28611 0.43828500000000004 0.989247\n",
      "28616 0.43828500000000004 0.991902\n",
      "28622 0.43828500000000004 0.954936\n",
      "28626 0.43828500000000004 0.994883\n",
      "28667 0.43828500000000004 0.991866\n",
      "28771 0.43828500000000004 0.9938\n",
      "29006 0.43828500000000004 0.987662\n",
      "29023 0.43828500000000004 0.992293\n",
      "29031 0.43828500000000004 0.982293\n",
      "29047 0.43828500000000004 0.986681\n",
      "29076 0.43828500000000004 0.982302\n",
      "29235 0.43828500000000004 0.999864\n",
      "29348 0.43828500000000004 0.975227\n",
      "29356 0.43828500000000004 0.991866\n",
      "29415 0.43828500000000004 0.991579\n",
      "29509 0.43828500000000004 0.995918\n",
      "29515 0.43828500000000004 0.998684\n",
      "29653 0.43828500000000004 0.973566\n",
      "29656 0.43828500000000004 0.999147\n",
      "29811 0.43828500000000004 0.95519\n",
      "29871 0.43828500000000004 0.990672\n",
      "29900 0.43828500000000004 0.99964\n",
      "30049 0.43828500000000004 0.972557\n",
      "30066 0.43828500000000004 0.996613\n",
      "30071 0.43828500000000004 0.952897\n",
      "30083 0.43828500000000004 0.988841\n",
      "30145 0.43828500000000004 0.989003\n",
      "30198 0.43828500000000004 0.97734\n",
      "30229 0.43828500000000004 0.981753\n",
      "30241 0.43828500000000004 0.965786\n",
      "30319 0.43828500000000004 0.992439\n",
      "30333 0.43828500000000004 0.995926\n",
      "30539 0.43828500000000004 0.982302\n",
      "30569 0.43828500000000004 0.975108\n",
      "30640 0.43828500000000004 0.991985\n",
      "30641 0.43828500000000004 0.993527\n",
      "30919 0.43828500000000004 0.991667\n",
      "30980 0.43828500000000004 0.964338\n",
      "31155 0.43828500000000004 0.985454\n",
      "31181 0.43828500000000004 0.965359\n",
      "31202 0.43828500000000004 0.982684\n",
      "31253 0.43828500000000004 0.987583\n",
      "31264 0.43828500000000004 0.974758\n",
      "31329 0.43828500000000004 0.9938\n",
      "31473 0.43828500000000004 0.954936\n",
      "31568 0.43828500000000004 0.952963\n",
      "31573 0.43828500000000004 0.991118\n",
      "31736 0.43828500000000004 0.988543\n",
      "31741 0.43828500000000004 0.991349\n",
      "31807 0.43828500000000004 0.978574\n",
      "31832 0.43828500000000004 0.994421\n",
      "31842 0.43828500000000004 0.991866\n",
      "31852 0.43828500000000004 0.993009\n",
      "31958 0.43828500000000004 0.957937\n",
      "32075 0.43828500000000004 1.0\n",
      "32093 0.43828500000000004 0.961557\n",
      "32095 0.43828500000000004 0.989003\n",
      "32112 0.43828500000000004 0.993983\n",
      "32233 0.43828500000000004 0.987184\n",
      "32348 0.43828500000000004 0.965359\n",
      "32382 0.43828500000000004 0.967767\n",
      "32388 0.43828500000000004 0.998537\n",
      "32402 0.43828500000000004 0.99837\n",
      "32449 0.43828500000000004 0.999021\n",
      "32591 0.43828500000000004 0.977482\n",
      "32702 0.43828500000000004 0.993343\n",
      "32750 0.43828500000000004 0.961135\n",
      "32849 0.43828500000000004 0.997999\n",
      "32876 0.43828500000000004 0.962726\n",
      "32880 0.43828500000000004 0.978156\n",
      "32966 0.43828500000000004 0.965011\n",
      "33126 0.43828500000000004 0.992804\n",
      "33155 0.43828500000000004 0.994097\n",
      "33170 0.43828500000000004 0.991706\n",
      "33186 0.43828500000000004 0.998292\n",
      "33199 0.43828500000000004 0.997553\n",
      "33255 0.43828500000000004 0.993009\n",
      "33287 0.43828500000000004 0.95519\n",
      "33345 0.43828500000000004 0.997435\n",
      "33479 0.43828500000000004 0.993724\n",
      "33515 0.43828500000000004 0.991681\n",
      "33618 0.43828500000000004 0.963687\n",
      "33694 0.43828500000000004 0.989659\n",
      "33770 0.43828500000000004 0.965948\n",
      "33805 0.43828500000000004 0.959017\n",
      "34004 0.43828500000000004 0.987583\n",
      "34077 0.43828500000000004 0.967843\n",
      "34114 0.43828500000000004 0.957033\n",
      "34379 0.43828500000000004 0.991667\n",
      "34430 0.43828500000000004 0.978156\n",
      "34504 0.43828500000000004 0.95851\n",
      "34698 0.43828500000000004 0.95519\n",
      "34778 0.43828500000000004 0.986778\n",
      "34907 0.43828500000000004 0.997098\n",
      "34908 0.43828500000000004 0.993343\n",
      "34995 0.43828500000000004 0.993938\n",
      "35038 0.43828500000000004 0.980642\n",
      "35052 0.43828500000000004 0.955169\n",
      "35086 0.43828500000000004 0.993261\n",
      "35127 0.43828500000000004 0.970141\n",
      "35138 0.43828500000000004 0.969344\n",
      "35193 0.43828500000000004 0.985426\n",
      "35257 0.43828500000000004 0.991302\n",
      "35355 0.43828500000000004 0.993009\n",
      "35433 0.43828500000000004 0.983131\n",
      "35450 0.43828500000000004 0.990343\n",
      "35527 0.43828500000000004 0.987553\n",
      "35564 0.43828500000000004 0.98473\n",
      "35684 0.43828500000000004 0.994791\n",
      "35890 0.43828500000000004 0.987184\n",
      "35937 0.43828500000000004 0.999996\n",
      "35950 0.43828500000000004 0.957937\n",
      "35998 0.43828500000000004 0.963078\n",
      "36087 0.43828500000000004 0.966972\n",
      "36104 0.43828500000000004 0.979276\n",
      "36157 0.43828500000000004 0.969344\n",
      "36184 0.43828500000000004 0.956727\n",
      "36245 0.43828500000000004 0.990853\n",
      "36265 0.43828500000000004 0.996309\n",
      "36379 0.43828500000000004 0.998444\n",
      "36412 0.43828500000000004 0.987184\n",
      "36418 0.43828500000000004 0.972119\n",
      "36420 0.43828500000000004 0.956727\n",
      "36437 0.43828500000000004 0.999021\n",
      "36461 0.43828500000000004 0.993909\n",
      "36500 0.43828500000000004 0.987184\n",
      "36513 0.43828500000000004 0.994678\n",
      "36592 0.43828500000000004 0.972728\n",
      "36662 0.43828500000000004 0.989331\n",
      "36721 0.43828500000000004 0.997479\n",
      "36794 0.43828500000000004 0.992198\n",
      "36921 0.43828500000000004 0.973566\n",
      "36925 0.43828500000000004 0.989659\n",
      "37016 0.43828500000000004 0.95519\n",
      "37030 0.43828500000000004 0.997553\n",
      "37054 0.43828500000000004 0.993343\n",
      "37059 0.43828500000000004 0.992804\n",
      "37159 0.43828500000000004 0.993536\n",
      "37168 0.43828500000000004 0.996613\n",
      "37223 0.43828500000000004 0.999021\n",
      "37230 0.43828500000000004 0.993009\n",
      "37259 0.43828500000000004 1.0\n",
      "37491 0.43828500000000004 0.994097\n",
      "37496 0.43828500000000004 0.976002\n",
      "37740 0.43828500000000004 0.989659\n",
      "37762 0.43828500000000004 0.956727\n",
      "37772 0.43828500000000004 0.965011\n",
      "37866 0.43828500000000004 0.997909\n",
      "38051 0.43828500000000004 0.95519\n",
      "38082 0.43828500000000004 0.965011\n",
      "38166 0.43828500000000004 0.998283\n",
      "38229 0.43828500000000004 0.991349\n",
      "38318 0.43828500000000004 0.992293\n",
      "38378 0.43828500000000004 0.999765\n",
      "38458 0.43828500000000004 0.98969\n",
      "38469 0.43828500000000004 0.97518\n",
      "38553 0.43828500000000004 0.989409\n",
      "38684 0.43828500000000004 0.999996\n",
      "38797 0.43828500000000004 0.991579\n",
      "38855 0.43828500000000004 0.984663\n",
      "38864 0.43828500000000004 0.983131\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "38893 0.43828500000000004 0.967843\n",
      "38968 0.43828500000000004 0.985574\n",
      "39054 0.43828500000000004 0.997984\n",
      "39185 0.43828500000000004 0.989014\n",
      "39256 0.43828500000000004 0.982293\n",
      "39372 0.43828500000000004 0.970127\n",
      "39411 0.43828500000000004 0.971222\n",
      "39483 0.43828500000000004 0.970369\n",
      "39540 0.43828500000000004 0.998292\n",
      "39563 0.43828500000000004 0.970369\n",
      "39570 0.43828500000000004 0.973566\n",
      "39579 0.43828500000000004 0.991866\n",
      "39696 0.43828500000000004 0.979693\n",
      "39811 0.43828500000000004 0.954936\n",
      "39873 0.43828500000000004 0.976002\n",
      "39892 0.43828500000000004 0.971264\n",
      "39953 0.43828500000000004 0.956727\n",
      "39999 0.43828500000000004 0.973705\n",
      "40090 0.43828500000000004 0.989409\n",
      "40095 0.43828500000000004 0.969197\n",
      "40123 0.43828500000000004 0.989659\n",
      "40141 0.43828500000000004 0.987553\n",
      "40160 0.43828500000000004 0.953622\n",
      "40246 0.43828500000000004 0.96192\n",
      "40418 0.43828500000000004 0.958451\n",
      "40486 0.43828500000000004 0.986583\n",
      "40514 0.43828500000000004 0.970749\n",
      "40530 0.43828500000000004 0.957738\n",
      "40604 0.43828500000000004 0.971222\n",
      "40705 0.43828500000000004 0.972119\n",
      "40718 0.43828500000000004 0.994421\n",
      "40817 0.43828500000000004 0.960541\n",
      "40928 0.43828500000000004 0.952897\n",
      "40964 0.43828500000000004 0.982302\n",
      "40990 0.43828500000000004 0.972119\n",
      "41072 0.43828500000000004 0.998909\n",
      "41084 0.43828500000000004 0.987184\n",
      "41164 0.43828500000000004 0.974514\n",
      "41287 0.43828500000000004 0.999021\n",
      "41306 0.43828500000000004 0.978156\n",
      "41316 0.43828500000000004 0.996718\n",
      "41446 0.43828500000000004 0.993009\n",
      "41451 0.43828500000000004 0.99462\n",
      "41552 0.43828500000000004 0.972596\n",
      "41554 0.43828500000000004 0.992926\n",
      "41567 0.43828500000000004 0.970749\n",
      "41621 0.43828500000000004 0.952897\n",
      "41697 0.43828500000000004 0.970749\n",
      "41767 0.43828500000000004 0.95519\n",
      "41781 0.43828500000000004 0.976426\n",
      "41870 0.43828500000000004 0.959017\n",
      "41956 0.43828500000000004 0.972119\n",
      "42145 0.43828500000000004 0.970693\n",
      "42187 0.43828500000000004 0.989409\n",
      "42202 0.43828500000000004 0.954936\n",
      "42215 0.43828500000000004 0.95519\n",
      "42341 0.43828500000000004 0.987184\n",
      "42348 0.43828500000000004 0.99873\n",
      "42376 0.43828500000000004 0.98969\n",
      "42388 0.43828500000000004 0.999708\n",
      "42456 0.43828500000000004 0.965948\n",
      "42517 0.43828500000000004 0.999259\n",
      "42544 0.43828500000000004 0.989003\n",
      "42731 0.43828500000000004 0.961135\n",
      "42765 0.43828500000000004 0.999568\n",
      "42861 0.43828500000000004 0.98473\n",
      "42879 0.43828500000000004 0.991579\n",
      "42884 0.43828500000000004 0.967843\n",
      "43068 0.43828500000000004 0.966972\n",
      "43440 0.43828500000000004 0.978156\n",
      "43562 0.43828500000000004 0.960541\n",
      "43627 0.43828500000000004 0.993837\n",
      "43690 0.43828500000000004 0.998444\n",
      "43709 0.43828500000000004 0.98473\n",
      "43726 0.43828500000000004 0.99805\n",
      "43730 0.43828500000000004 0.962637\n",
      "43911 0.43828500000000004 0.987553\n",
      "43959 0.43828500000000004 0.991866\n",
      "44070 0.43828500000000004 0.998292\n",
      "44429 0.43828500000000004 0.991579\n",
      "44503 0.43828500000000004 0.998257\n",
      "44613 0.43828500000000004 0.95559\n",
      "44650 0.43828500000000004 0.983131\n",
      "44661 0.43828500000000004 0.95519\n",
      "44696 0.43828500000000004 0.998671\n",
      "44725 0.43828500000000004 0.992293\n",
      "44761 0.43828500000000004 0.988841\n",
      "44841 0.43828500000000004 0.952963\n",
      "44925 0.43828500000000004 0.970917\n",
      "45016 0.43828500000000004 0.993405\n",
      "45095 0.43828500000000004 0.998257\n",
      "45262 0.43828500000000004 0.997553\n",
      "45367 0.43828500000000004 0.969344\n",
      "45507 0.43828500000000004 0.9938\n",
      "45662 0.43828500000000004 0.993571\n",
      "45723 0.43828500000000004 0.965425\n",
      "45738 0.43828500000000004 0.988543\n",
      "45898 0.43828500000000004 0.973894\n",
      "45922 0.43828500000000004 0.956727\n",
      "45957 0.43828500000000004 0.961135\n",
      "46029 0.43828500000000004 0.997999\n",
      "46061 0.43828500000000004 0.962637\n",
      "46099 0.43828500000000004 0.997984\n",
      "46106 0.43828500000000004 0.994888\n",
      "46113 0.43828500000000004 0.980829\n",
      "46159 0.43828500000000004 0.998283\n",
      "46170 0.43828500000000004 0.965425\n",
      "46213 0.43828500000000004 0.973001\n",
      "46232 0.43828500000000004 0.989659\n",
      "46233 0.43828500000000004 0.962637\n",
      "46287 0.43828500000000004 0.991349\n",
      "46330 0.43828500000000004 0.972119\n",
      "46414 0.43828500000000004 0.95483\n",
      "46713 0.43828500000000004 0.965359\n",
      "46722 0.43828500000000004 0.960335\n",
      "46778 0.43828500000000004 0.985838\n",
      "46794 0.43828500000000004 0.954509\n",
      "46873 0.43828500000000004 0.999021\n",
      "46992 0.43828500000000004 0.997693\n",
      "47001 0.43828500000000004 0.999021\n",
      "47012 0.43828500000000004 0.99837\n",
      "47082 0.43828500000000004 0.962726\n",
      "47408 0.43828500000000004 0.99837\n",
      "47442 0.43828500000000004 0.987184\n",
      "47531 0.43828500000000004 0.991579\n",
      "47687 0.43828500000000004 0.986386\n",
      "47715 0.43828500000000004 0.961135\n",
      "47735 0.43828500000000004 0.970917\n",
      "47766 0.43828500000000004 0.990026\n",
      "47820 0.43828500000000004 0.992786\n",
      "47825 0.43828500000000004 0.992293\n",
      "47868 0.43828500000000004 0.983623\n",
      "47904 0.43828500000000004 0.980539\n",
      "48000 0.43828500000000004 0.977514\n",
      "48093 0.43828500000000004 0.988543\n",
      "48250 0.43828500000000004 0.984117\n",
      "48327 0.43828500000000004 0.957033\n",
      "48595 0.43828500000000004 0.979693\n",
      "48597 0.43828500000000004 0.996527\n",
      "48695 0.43828500000000004 0.979436\n",
      "48937 0.43828500000000004 0.984826\n",
      "48955 0.43828500000000004 0.975944\n",
      "49214 0.43828500000000004 0.999021\n",
      "49319 0.43828500000000004 0.993114\n",
      "49367 0.43828500000000004 0.997553\n",
      "49546 0.43828500000000004 0.990695\n",
      "49591 0.43828500000000004 0.99837\n",
      "49660 0.43828500000000004 0.979585\n",
      "49691 0.43828500000000004 0.970369\n",
      "49712 0.43828500000000004 0.95519\n",
      "49804 0.43828500000000004 0.964202\n",
      "49879 0.43828500000000004 0.972119\n",
      "49883 0.43828500000000004 0.95519\n",
      "50023 0.43828500000000004 0.989409\n",
      "50265 0.43828500000000004 0.998101\n",
      "50519 0.43828500000000004 0.995762\n",
      "50559 0.43828500000000004 0.970749\n",
      "50563 0.43828500000000004 0.996808\n",
      "50590 0.43828500000000004 0.988543\n",
      "50685 0.43828500000000004 0.967767\n",
      "50811 0.43828500000000004 0.957937\n",
      "50855 0.43828500000000004 0.986803\n",
      "50886 0.43828500000000004 0.997909\n",
      "50959 0.43828500000000004 0.999095\n",
      "50989 0.43828500000000004 0.988543\n",
      "51113 0.43828500000000004 0.972119\n",
      "51114 0.43828500000000004 0.991902\n",
      "51137 0.43828500000000004 0.986778\n",
      "51266 0.43828500000000004 0.98473\n",
      "51541 0.43828500000000004 0.987184\n",
      "51577 0.43828500000000004 0.986803\n",
      "51583 0.43828500000000004 0.993343\n",
      "51662 0.43828500000000004 0.999783\n",
      "51761 0.43828500000000004 0.997553\n",
      "51795 0.43828500000000004 0.997553\n",
      "51797 0.43828500000000004 0.952897\n",
      "51817 0.43828500000000004 0.993724\n",
      "51938 0.43828500000000004 0.970749\n",
      "52021 0.43828500000000004 0.994678\n",
      "52037 0.43828500000000004 0.980958\n",
      "52125 0.43828500000000004 0.998854\n",
      "52211 0.43828500000000004 0.95519\n",
      "52228 0.43828500000000004 0.990343\n",
      "52239 0.43828500000000004 0.989659\n",
      "52313 0.43828500000000004 0.979276\n",
      "52333 0.43828500000000004 0.95519\n",
      "52435 0.43828500000000004 0.979761\n",
      "52490 0.43828500000000004 0.990343\n",
      "52572 0.43828500000000004 0.994396\n",
      "52594 0.43828500000000004 0.997553\n",
      "52603 0.43828500000000004 0.970917\n",
      "52630 0.43828500000000004 0.991479\n",
      "52859 0.43828500000000004 0.989659\n",
      "52885 0.43828500000000004 0.99336\n",
      "52902 0.43828500000000004 0.985969\n",
      "53020 0.43828500000000004 0.953228\n",
      "53116 0.43828500000000004 0.960354\n",
      "53239 0.43828500000000004 0.952963\n",
      "53278 0.43828500000000004 0.970917\n",
      "53312 0.43828500000000004 0.999021\n",
      "53329 0.43828500000000004 0.95559\n",
      "53362 0.43828500000000004 0.993909\n",
      "53364 0.43828500000000004 0.99837\n",
      "53381 0.43828500000000004 0.988543\n",
      "53417 0.43828500000000004 0.987516\n",
      "53422 0.43828500000000004 0.962637\n",
      "53428 0.43828500000000004 0.985735\n",
      "53469 0.43828500000000004 0.983131\n",
      "53551 0.43828500000000004 1.0\n",
      "53701 0.43828500000000004 0.998444\n",
      "53781 0.43828500000000004 0.965425\n",
      "53888 0.43828500000000004 0.998854\n",
      "53965 0.43828500000000004 0.990343\n",
      "54031 0.43828500000000004 0.997553\n",
      "54069 0.43828500000000004 0.986583\n",
      "54099 0.43828500000000004 0.973705\n",
      "54188 0.43828500000000004 0.990695\n",
      "54201 0.43828500000000004 0.999973\n",
      "54338 0.43828500000000004 0.996931\n",
      "54376 0.43828500000000004 0.997553\n",
      "54413 0.43828500000000004 0.966972\n",
      "54570 0.43828500000000004 0.987184\n",
      "54673 0.43828500000000004 0.95483\n",
      "54692 0.43828500000000004 0.999573\n",
      "54697 0.43828500000000004 0.991886\n",
      "54715 0.43828500000000004 0.976141\n",
      "54885 0.43828500000000004 0.999708\n",
      "54896 0.43828500000000004 0.975653\n",
      "54922 0.43828500000000004 0.991667\n",
      "54952 0.43828500000000004 0.954316\n",
      "54968 0.43828500000000004 0.986386\n",
      "55180 0.43828500000000004 0.965359\n",
      "55294 0.43828500000000004 0.979585\n",
      "55338 0.43828500000000004 0.962867\n",
      "55697 0.43828500000000004 0.95559\n",
      "55729 0.43828500000000004 0.956727\n",
      "55738 0.43828500000000004 0.971222\n",
      "55930 0.43828500000000004 0.95519\n",
      "55961 0.43828500000000004 0.967767\n",
      "55977 0.43828500000000004 0.979693\n",
      "56018 0.43828500000000004 0.972596\n",
      "56064 0.43828500000000004 0.973489\n",
      "56082 0.43828500000000004 0.992926\n",
      "56123 0.43828500000000004 0.999147\n",
      "56148 0.43828500000000004 0.991554\n",
      "56201 0.43828500000000004 0.99462\n",
      "56254 0.43828500000000004 0.958353\n",
      "56270 0.43828500000000004 0.997553\n",
      "56319 0.43828500000000004 0.962637\n",
      "56419 0.43828500000000004 0.997553\n",
      "56486 0.43828500000000004 0.967843\n",
      "56599 0.43828500000000004 0.994097\n",
      "56619 0.43828500000000004 0.996527\n",
      "56655 0.43828500000000004 0.970917\n",
      "56668 0.43828500000000004 0.973489\n",
      "56689 0.43828500000000004 0.998161\n",
      "56696 0.43828500000000004 0.994696\n",
      "56774 0.43828500000000004 0.999568\n",
      "56813 0.43828500000000004 0.95519\n",
      "56877 0.43828500000000004 0.972557\n",
      "56907 0.43828500000000004 0.985762\n",
      "56956 0.43828500000000004 0.990343\n",
      "57006 0.43828500000000004 0.972119\n",
      "57010 0.43828500000000004 0.99433\n",
      "57074 0.43828500000000004 0.993343\n",
      "57084 0.43828500000000004 1.0\n",
      "57113 0.43828500000000004 0.995316\n",
      "57233 0.43828500000000004 0.998537\n",
      "57289 0.43828500000000004 0.992293\n",
      "57408 0.43828500000000004 0.95519\n",
      "57424 0.43828500000000004 0.990736\n",
      "57431 0.43828500000000004 0.989331\n",
      "57472 0.43828500000000004 0.989003\n",
      "57531 0.43828500000000004 0.994421\n",
      "57533 0.43828500000000004 0.999975\n",
      "57550 0.43828500000000004 0.987184\n",
      "57603 0.43828500000000004 0.997553\n",
      "57606 0.43828500000000004 0.96192\n",
      "57705 0.43828500000000004 0.96521\n",
      "57711 0.43828500000000004 0.961557\n",
      "57718 0.43828500000000004 0.95483\n",
      "57721 0.43828500000000004 0.996647\n",
      "57835 0.43828500000000004 0.998283\n",
      "57858 0.43828500000000004 0.994492\n",
      "57936 0.43828500000000004 0.989424\n",
      "58008 0.43828500000000004 0.997553\n",
      "58125 0.43828500000000004 0.95519\n",
      "58173 0.43828500000000004 0.957033\n",
      "58189 0.43828500000000004 0.973566\n",
      "58278 0.43828500000000004 0.966972\n",
      "58336 0.43828500000000004 0.965622\n",
      "58394 0.43828500000000004 0.983464\n",
      "58395 0.43828500000000004 0.989409\n",
      "58455 0.43828500000000004 0.989659\n",
      "58592 0.43828500000000004 0.995762\n",
      "58649 0.43828500000000004 0.960071\n",
      "58664 0.43828500000000004 0.960541\n",
      "58744 0.43828500000000004 0.960071\n",
      "58776 0.43828500000000004 0.978103\n",
      "59115 0.43828500000000004 0.957937\n",
      "59195 0.43828500000000004 0.95519\n",
      "59274 0.43828500000000004 0.987184\n",
      "59324 0.43828500000000004 0.997734\n",
      "59440 0.43828500000000004 0.995723\n",
      "59513 0.43828500000000004 0.985762\n",
      "59525 0.43828500000000004 0.992293\n",
      "59597 0.43828500000000004 0.957937\n",
      "59723 0.43828500000000004 0.997734\n",
      "59799 0.43828500000000004 0.999573\n",
      "59844 0.43828500000000004 0.993565\n",
      "59885 0.43828500000000004 0.960541\n",
      "59938 0.43828500000000004 0.996647\n",
      "60001 0.43828500000000004 0.952963\n",
      "60018 0.43828500000000004 0.95519\n",
      "60034 0.43828500000000004 0.994097\n",
      "60133 0.43828500000000004 0.985735\n",
      "60261 0.43828500000000004 0.970749\n",
      "60485 0.43828500000000004 0.992676\n",
      "60633 0.43828500000000004 0.990853\n",
      "60657 0.43828500000000004 0.965011\n",
      "60665 0.43828500000000004 0.952897\n",
      "60713 0.43828500000000004 0.988543\n",
      "60749 0.43828500000000004 0.98498\n",
      "60765 0.43828500000000004 0.954316\n",
      "61054 0.43828500000000004 0.953228\n",
      "61140 0.43828500000000004 0.960335\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "61150 0.43828500000000004 0.998292\n",
      "61184 0.43828500000000004 0.986771\n",
      "61198 0.43828500000000004 0.987184\n",
      "61220 0.43828500000000004 0.997339\n",
      "61467 0.43828500000000004 0.972119\n",
      "61483 0.43828500000000004 0.993009\n",
      "61522 0.43828500000000004 0.997734\n",
      "61582 0.43828500000000004 0.989659\n",
      "61772 0.43828500000000004 0.965044\n",
      "61900 0.43828500000000004 0.968467\n",
      "61909 0.43828500000000004 0.958451\n",
      "61914 0.43828500000000004 0.982293\n",
      "62058 0.43828500000000004 0.984558\n",
      "62174 0.43828500000000004 0.986583\n",
      "62349 0.43828500000000004 0.982851\n",
      "62426 0.43828500000000004 0.982302\n",
      "62514 0.43828500000000004 0.991349\n",
      "62537 0.43828500000000004 0.99837\n",
      "62562 0.43828500000000004 0.969344\n",
      "62605 0.43828500000000004 0.990736\n",
      "62636 0.43828500000000004 0.993114\n",
      "62653 0.43828500000000004 0.970749\n",
      "62687 0.43828500000000004 0.996334\n",
      "62865 0.43828500000000004 0.987184\n",
      "62889 0.43828500000000004 0.987033\n",
      "62967 0.43828500000000004 0.965011\n",
      "62976 0.43828500000000004 0.980539\n",
      "62990 0.43828500000000004 0.965786\n",
      "63034 0.43828500000000004 0.979761\n",
      "63208 0.43828500000000004 0.981806\n",
      "63225 0.43828500000000004 0.991866\n",
      "63272 0.43828500000000004 0.973604\n",
      "63352 0.43828500000000004 0.990343\n",
      "63421 0.43828500000000004 0.987184\n",
      "63483 0.43828500000000004 0.990853\n",
      "63527 0.43828500000000004 0.99846\n",
      "63857 0.43828500000000004 0.959423\n",
      "63858 0.43828500000000004 0.975944\n",
      "63958 0.43828500000000004 0.989948\n",
      "63964 0.43828500000000004 0.995762\n",
      "64012 0.43828500000000004 0.960429\n",
      "64053 0.43828500000000004 1.0\n",
      "64112 0.43828500000000004 0.957937\n",
      "64276 0.43828500000000004 0.991349\n",
      "64280 0.43828500000000004 0.988841\n",
      "64517 0.43828500000000004 0.985762\n",
      "64679 0.43828500000000004 0.972119\n",
      "64868 0.43828500000000004 0.95519\n",
      "64873 0.43828500000000004 0.980829\n",
      "64974 0.43828500000000004 0.960335\n",
      "65024 0.43828500000000004 0.987553\n",
      "65037 0.43828500000000004 0.993983\n",
      "65177 0.43828500000000004 0.978156\n",
      "65229 0.43828500000000004 0.998073\n",
      "65246 0.43828500000000004 0.985454\n",
      "65452 0.43828500000000004 0.992786\n",
      "65479 0.43828500000000004 0.982293\n",
      "65486 0.43828500000000004 0.986681\n",
      "65515 0.43828500000000004 0.973604\n",
      "65542 0.43828500000000004 0.997909\n",
      "65577 0.43828500000000004 0.999021\n",
      "65765 0.43828500000000004 0.997999\n",
      "65809 0.43828500000000004 0.995845\n",
      "65859 0.43828500000000004 0.998292\n",
      "65950 0.43828500000000004 0.996914\n",
      "65957 0.43828500000000004 0.952963\n",
      "66194 0.43828500000000004 0.998444\n",
      "66236 0.43828500000000004 0.99858\n",
      "66382 0.43828500000000004 0.99873\n",
      "66497 0.43828500000000004 0.990343\n",
      "66572 0.43828500000000004 0.993724\n",
      "66597 0.43828500000000004 0.977482\n",
      "66625 0.43828500000000004 0.972596\n",
      "66784 0.43828500000000004 0.966639\n",
      "66795 0.43828500000000004 0.95483\n",
      "66823 0.43828500000000004 0.996173\n",
      "66843 0.43828500000000004 0.997734\n",
      "66935 0.43828500000000004 0.984377\n",
      "66992 0.43828500000000004 0.993983\n",
      "67053 0.43828500000000004 0.994845\n",
      "67108 0.43828500000000004 0.993009\n",
      "67228 0.43828500000000004 0.996094\n",
      "67322 0.43828500000000004 0.976002\n",
      "67356 0.43828500000000004 0.983464\n",
      "67407 0.43828500000000004 0.987184\n",
      "67453 0.43828500000000004 0.997761\n",
      "67455 0.43828500000000004 0.973705\n",
      "67480 0.43828500000000004 0.998292\n",
      "67533 0.43828500000000004 0.965359\n",
      "67622 0.43828500000000004 0.952519\n",
      "67698 0.43828500000000004 0.989424\n",
      "67710 0.43828500000000004 0.953775\n",
      "67726 0.43828500000000004 0.997984\n",
      "67744 0.43828500000000004 0.95559\n",
      "67783 0.43828500000000004 0.96192\n",
      "67926 0.43828500000000004 0.980539\n",
      "67939 0.43828500000000004 0.989003\n",
      "67983 0.43828500000000004 0.973566\n",
      "68066 0.43828500000000004 0.974758\n",
      "68156 0.43828500000000004 0.968012\n",
      "68228 0.43828500000000004 0.970917\n",
      "68252 0.43828500000000004 0.971222\n",
      "68379 0.43828500000000004 0.996908\n",
      "68495 0.43828500000000004 0.986386\n",
      "68534 0.43828500000000004 0.966972\n",
      "68578 0.43828500000000004 0.98969\n",
      "68590 0.43828500000000004 0.954936\n",
      "68832 0.43828500000000004 0.999095\n",
      "68904 0.43828500000000004 0.964338\n",
      "69000 0.43828500000000004 0.962637\n",
      "69075 0.43828500000000004 0.994421\n",
      "69190 0.43828500000000004 0.96192\n",
      "69403 0.43828500000000004 0.985735\n",
      "69633 0.43828500000000004 0.988543\n",
      "69668 0.43828500000000004 0.965011\n",
      "69719 0.43828500000000004 0.988841\n",
      "69806 0.43828500000000004 0.972119\n",
      "69851 0.43828500000000004 0.999021\n",
      "69871 0.43828500000000004 0.96192\n",
      "69880 0.43828500000000004 0.984745\n",
      "70176 0.43828500000000004 0.969998\n",
      "70286 0.43828500000000004 0.980642\n",
      "70332 0.43828500000000004 0.959423\n",
      "70413 0.43828500000000004 0.962726\n",
      "70414 0.43828500000000004 0.984973\n",
      "70505 0.43828500000000004 0.972596\n",
      "70508 0.43828500000000004 0.971264\n",
      "70591 0.43828500000000004 0.982293\n",
      "70642 0.43828500000000004 0.975653\n",
      "70660 0.43828500000000004 0.980091\n",
      "70982 0.43828500000000004 0.993527\n",
      "70996 0.43828500000000004 0.996553\n",
      "71159 0.43828500000000004 0.954936\n",
      "71161 0.43828500000000004 0.954531\n",
      "71182 0.43828500000000004 0.962867\n",
      "71263 0.43828500000000004 0.998537\n",
      "71353 0.43828500000000004 0.991667\n",
      "71367 0.43828500000000004 0.997704\n",
      "71486 0.43828500000000004 0.997553\n",
      "71588 0.43828500000000004 0.957033\n",
      "71639 0.43828500000000004 0.965011\n",
      "71679 0.43828500000000004 0.989659\n",
      "71769 0.43828500000000004 0.98473\n",
      "71790 0.43828500000000004 0.992293\n",
      "71878 0.43828500000000004 0.999827\n",
      "71888 0.43828500000000004 0.981753\n",
      "72024 0.43828500000000004 0.980958\n",
      "72076 0.43828500000000004 1.0\n",
      "72085 0.43828500000000004 0.993938\n",
      "72257 0.43828500000000004 0.970917\n",
      "72345 0.43828500000000004 0.981023\n",
      "72374 0.43828500000000004 0.974514\n",
      "72525 0.43828500000000004 0.973566\n",
      "72558 0.43828500000000004 0.996247\n",
      "72581 0.43828500000000004 0.998811\n",
      "72593 0.43828500000000004 0.9998\n",
      "72600 0.43828500000000004 0.989659\n",
      "72616 0.43828500000000004 0.997553\n",
      "72733 0.43828500000000004 0.997299\n",
      "72934 0.43828500000000004 0.975227\n",
      "73066 0.43828500000000004 0.972119\n",
      "73077 0.43828500000000004 0.975653\n",
      "73163 0.43828500000000004 0.99792\n",
      "73335 0.43828500000000004 0.980539\n",
      "73351 0.43828500000000004 0.975227\n",
      "73445 0.43828500000000004 0.970764\n",
      "73448 0.43828500000000004 0.997429\n",
      "73484 0.43828500000000004 0.965359\n",
      "73579 0.43828500000000004 0.960071\n",
      "73781 0.43828500000000004 0.996718\n",
      "73877 0.43828500000000004 0.980539\n",
      "74201 0.43828500000000004 0.996914\n",
      "74305 0.43828500000000004 0.967767\n",
      "74308 0.43828500000000004 0.990296\n",
      "74606 0.43828500000000004 0.991866\n",
      "74850 0.43828500000000004 0.985762\n",
      "74903 0.43828500000000004 0.970749\n",
      "74916 0.43828500000000004 0.95519\n",
      "75079 0.43828500000000004 0.985735\n",
      "75115 0.43828500000000004 0.953622\n",
      "75120 0.43828500000000004 0.95483\n",
      "75140 0.43828500000000004 0.963687\n",
      "75142 0.43828500000000004 0.997999\n",
      "75180 0.43828500000000004 0.984826\n",
      "75190 0.43828500000000004 0.965011\n",
      "75417 0.43828500000000004 0.985248\n",
      "75512 0.43828500000000004 0.970749\n",
      "75519 0.43828500000000004 0.960071\n",
      "75530 0.43828500000000004 0.994421\n",
      "75551 0.43828500000000004 0.962726\n",
      "75651 0.43828500000000004 0.998551\n",
      "75754 0.43828500000000004 0.996718\n",
      "75785 0.43828500000000004 0.95483\n",
      "75855 0.43828500000000004 0.970917\n",
      "75916 0.43828500000000004 0.985735\n",
      "75997 0.43828500000000004 0.994545\n",
      "76059 0.43828500000000004 0.9969\n",
      "76122 0.43828500000000004 0.990853\n",
      "76146 0.43828500000000004 0.957033\n",
      "76248 0.43828500000000004 0.972119\n",
      "76394 0.43828500000000004 0.952897\n",
      "76415 0.43828500000000004 0.965425\n",
      "76432 0.43828500000000004 0.975712\n",
      "76495 0.43828500000000004 0.994994\n",
      "76780 0.43828500000000004 0.997553\n",
      "76817 0.43828500000000004 0.978156\n",
      "76962 0.43828500000000004 0.997847\n",
      "77117 0.43828500000000004 0.989247\n",
      "77178 0.43828500000000004 0.997339\n",
      "77242 0.43828500000000004 0.984826\n",
      "77462 0.43828500000000004 0.979585\n",
      "77549 0.43828500000000004 0.976652\n",
      "77592 0.43828500000000004 0.95519\n",
      "77664 0.43828500000000004 0.979585\n",
      "77669 0.43828500000000004 0.997693\n",
      "77734 0.43828500000000004 0.994212\n",
      "77748 0.43828500000000004 0.994678\n",
      "77763 0.43828500000000004 0.999095\n",
      "77831 0.43828500000000004 0.975828\n",
      "77946 0.43828500000000004 0.970749\n",
      "77983 0.43828500000000004 0.961135\n",
      "78187 0.43828500000000004 0.994883\n",
      "78261 0.43828500000000004 0.959017\n",
      "78322 0.43828500000000004 0.98185\n",
      "78350 0.43828500000000004 0.992804\n",
      "78674 0.43828500000000004 0.954936\n",
      "78751 0.43828500000000004 0.965425\n",
      "78803 0.43828500000000004 0.999997\n",
      "78808 0.43828500000000004 0.987583\n",
      "78837 0.43828500000000004 0.99964\n",
      "78849 0.43828500000000004 0.965359\n",
      "78911 0.43828500000000004 0.973566\n",
      "78949 0.43828500000000004 0.970369\n",
      "78981 0.43828500000000004 0.960071\n",
      "79000 0.43828500000000004 0.954531\n",
      "79186 0.43828500000000004 0.972596\n",
      "79197 0.43828500000000004 0.980178\n",
      "79249 0.43828500000000004 0.989948\n",
      "79258 0.43828500000000004 0.965786\n",
      "79332 0.43828500000000004 0.95519\n",
      "79431 0.43828500000000004 0.998537\n",
      "79437 0.43828500000000004 0.970693\n",
      "79446 0.43828500000000004 0.978156\n",
      "79645 0.43828500000000004 0.95851\n",
      "79731 0.43828500000000004 0.960541\n",
      "79780 0.43828500000000004 0.993009\n",
      "79817 0.43828500000000004 0.965011\n",
      "79854 0.43828500000000004 0.993837\n",
      "80041 0.43828500000000004 0.994421\n",
      "80148 0.43828500000000004 0.95483\n",
      "80185 0.43828500000000004 0.960071\n",
      "80264 0.43828500000000004 0.994545\n",
      "80320 0.43828500000000004 0.993508\n",
      "80360 0.43828500000000004 0.981206\n",
      "80363 0.43828500000000004 0.970749\n",
      "80367 0.43828500000000004 0.992439\n",
      "80418 0.43828500000000004 0.994097\n",
      "80562 0.43828500000000004 0.969344\n",
      "80681 0.43828500000000004 0.999259\n",
      "80730 0.43828500000000004 0.994421\n",
      "80797 0.43828500000000004 0.962726\n",
      "80822 0.43828500000000004 0.994888\n",
      "80912 0.43828500000000004 0.998537\n",
      "80937 0.43828500000000004 0.994097\n",
      "80948 0.43828500000000004 0.983464\n",
      "81085 0.43828500000000004 0.958135\n",
      "81100 0.43828500000000004 0.965786\n",
      "81107 0.43828500000000004 0.987553\n",
      "81237 0.43828500000000004 0.992229\n",
      "81345 0.43828500000000004 0.980642\n",
      "81380 0.43828500000000004 0.977482\n",
      "81461 0.43828500000000004 0.991645\n",
      "81475 0.43828500000000004 0.999021\n",
      "81758 0.43828500000000004 0.97518\n",
      "81799 0.43828500000000004 0.984826\n",
      "81936 0.43828500000000004 0.985762\n",
      "82190 0.43828500000000004 0.973566\n",
      "82222 0.43828500000000004 0.988145\n",
      "82344 0.43828500000000004 0.998909\n",
      "82366 0.43828500000000004 0.989659\n",
      "82411 0.43828500000000004 0.994421\n",
      "82512 0.43828500000000004 0.982302\n",
      "82681 0.43828500000000004 0.998367\n",
      "82789 0.43828500000000004 0.969344\n",
      "82805 0.43828500000000004 0.965011\n",
      "82872 0.43828500000000004 0.992368\n",
      "83031 0.43828500000000004 0.987184\n",
      "83082 0.43828500000000004 0.987662\n",
      "83151 0.43828500000000004 0.993539\n",
      "83210 0.43828500000000004 0.991438\n",
      "83241 0.43828500000000004 0.99336\n",
      "83249 0.43828500000000004 0.984745\n",
      "83290 0.43828500000000004 0.997734\n",
      "83526 0.43828500000000004 0.985735\n",
      "83557 0.43828500000000004 0.979276\n",
      "83720 0.43828500000000004 0.991667\n",
      "83832 0.43828500000000004 0.952897\n",
      "83871 0.43828500000000004 1.0\n",
      "83890 0.43828500000000004 0.994143\n",
      "83982 0.43828500000000004 0.964117\n",
      "84187 0.43828500000000004 0.997553\n",
      "84295 0.43828500000000004 0.965359\n",
      "84323 0.43828500000000004 0.972119\n",
      "84402 0.43828500000000004 0.997909\n",
      "84413 0.43828500000000004 0.971222\n",
      "84470 0.43828500000000004 0.960541\n",
      "84475 0.43828500000000004 0.990704\n",
      "84636 0.43828500000000004 0.969344\n",
      "84706 0.43828500000000004 0.999999\n",
      "84857 0.43828500000000004 0.990343\n",
      "84926 0.43828500000000004 0.971606\n",
      "84951 0.43828500000000004 0.989003\n",
      "85017 0.43828500000000004 0.987553\n",
      "85021 0.43828500000000004 0.997704\n",
      "85169 0.43828500000000004 0.95483\n",
      "85182 0.43828500000000004 0.998292\n",
      "85193 0.43828500000000004 0.992676\n",
      "85267 0.43828500000000004 0.994678\n",
      "85341 0.43828500000000004 0.979436\n",
      "85342 0.43828500000000004 0.972596\n",
      "85364 0.43828500000000004 0.97577\n",
      "85369 0.43828500000000004 0.992439\n",
      "85444 0.43828500000000004 0.978923\n",
      "85511 0.43828500000000004 0.994883\n",
      "85593 0.43828500000000004 0.988543\n",
      "85646 0.43828500000000004 0.999573\n",
      "85805 0.43828500000000004 0.989973\n",
      "85880 0.43828500000000004 0.99462\n",
      "85919 0.43828500000000004 0.986681\n",
      "86118 0.43828500000000004 0.997734\n",
      "86140 0.43828500000000004 0.989659\n",
      "86215 0.43828500000000004 0.994532\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "86269 0.43828500000000004 0.974758\n",
      "86413 0.43828500000000004 0.989424\n",
      "86504 0.43828500000000004 0.993938\n",
      "86515 0.43828500000000004 0.991579\n",
      "86564 0.43828500000000004 0.96192\n",
      "86625 0.43828500000000004 0.973705\n",
      "86654 0.43828500000000004 0.993009\n",
      "86658 0.43828500000000004 0.972806\n",
      "86804 0.43828500000000004 0.987553\n",
      "86857 0.43828500000000004 0.970749\n",
      "86869 0.43828500000000004 0.972119\n",
      "86877 0.43828500000000004 0.955799\n",
      "86933 0.43828500000000004 0.987184\n",
      "87069 0.43828500000000004 0.997051\n",
      "87117 0.43828500000000004 0.991562\n",
      "87184 0.43828500000000004 0.983464\n",
      "87246 0.43828500000000004 0.999493\n",
      "87294 0.43828500000000004 0.964541\n",
      "87391 0.43828500000000004 0.96192\n",
      "87441 0.43828500000000004 0.99871\n",
      "87725 0.43828500000000004 0.989659\n",
      "87768 0.43828500000000004 0.997553\n",
      "87851 0.43828500000000004 0.952519\n",
      "87935 0.43828500000000004 0.988151\n",
      "88025 0.43828500000000004 0.977482\n",
      "88232 0.43828500000000004 0.996718\n",
      "88294 0.43828500000000004 0.985728\n",
      "88338 0.43828500000000004 0.96192\n",
      "88521 0.43828500000000004 0.994883\n",
      "88535 0.43828500000000004 0.98969\n",
      "88588 0.43828500000000004 0.983464\n",
      "88750 0.43828500000000004 0.970369\n",
      "88873 0.43828500000000004 0.997479\n",
      "88879 0.43828500000000004 0.967767\n",
      "88885 0.43828500000000004 0.99871\n",
      "88991 0.43828500000000004 0.978574\n",
      "89094 0.43828500000000004 0.969197\n",
      "89250 0.43828500000000004 0.97518\n",
      "89260 0.43828500000000004 0.973705\n",
      "89263 0.43828500000000004 0.99336\n",
      "89305 0.43828500000000004 0.998146\n",
      "89454 0.43828500000000004 0.979693\n",
      "89471 0.43828500000000004 0.990343\n",
      "89524 0.43828500000000004 0.996613\n",
      "89584 0.43828500000000004 0.998444\n",
      "89655 0.43828500000000004 0.983464\n",
      "89702 0.43828500000000004 0.994678\n",
      "89756 0.43828500000000004 0.953884\n",
      "89842 0.43828500000000004 0.980539\n",
      "89925 0.43828500000000004 0.990343\n",
      "89981 0.43828500000000004 0.970917\n",
      "90063 0.43828500000000004 0.999147\n",
      "90090 0.43828500000000004 0.998444\n",
      "90157 0.43828500000000004 0.961557\n",
      "90237 0.43828500000000004 0.994678\n",
      "90287 0.43828500000000004 0.997504\n",
      "90355 0.43828500000000004 0.997184\n",
      "90421 0.43828500000000004 0.997734\n",
      "90425 0.43828500000000004 0.997553\n",
      "90476 0.43828500000000004 0.997553\n",
      "90515 0.43828500000000004 0.987184\n",
      "90656 0.43828500000000004 0.95519\n",
      "90903 0.43828500000000004 0.975944\n",
      "90994 0.43828500000000004 0.965786\n",
      "91017 0.43828500000000004 0.953884\n",
      "91023 0.43828500000000004 0.991866\n",
      "91028 0.43828500000000004 0.996553\n",
      "91224 0.43828500000000004 0.972936\n",
      "91349 0.43828500000000004 0.993938\n",
      "91712 0.43828500000000004 0.972119\n",
      "91976 0.43828500000000004 0.993724\n",
      "91979 0.43828500000000004 0.991579\n",
      "92110 0.43828500000000004 0.981206\n",
      "92351 0.43828500000000004 0.966972\n",
      "92427 0.43828500000000004 0.955169\n",
      "92489 0.43828500000000004 0.987184\n",
      "92525 0.43828500000000004 0.994421\n",
      "92530 0.43828500000000004 0.994143\n",
      "92549 0.43828500000000004 0.984745\n",
      "92584 0.43828500000000004 0.987184\n",
      "92752 0.43828500000000004 0.993527\n",
      "92759 0.43828500000000004 0.992804\n",
      "92793 0.43828500000000004 0.987516\n",
      "93050 0.43828500000000004 0.986681\n",
      "93386 0.43828500000000004 1.0\n",
      "93430 0.43828500000000004 0.972119\n",
      "93437 0.43828500000000004 0.989331\n",
      "93526 0.43828500000000004 0.987184\n",
      "93757 0.43828500000000004 0.999896\n",
      "93856 0.43828500000000004 0.988841\n",
      "93938 0.43828500000000004 0.994097\n",
      "93942 0.43828500000000004 0.98498\n",
      "94022 0.43828500000000004 0.992676\n",
      "94083 0.43828500000000004 0.990704\n",
      "94130 0.43828500000000004 0.997339\n",
      "94133 0.43828500000000004 0.960429\n",
      "94157 0.43828500000000004 0.95483\n",
      "94184 0.43828500000000004 0.987583\n",
      "94202 0.43828500000000004 0.999095\n",
      "94304 0.43828500000000004 0.953884\n",
      "94317 0.43828500000000004 0.979585\n",
      "94363 0.43828500000000004 0.956727\n",
      "94544 0.43828500000000004 0.997504\n",
      "94622 0.43828500000000004 0.999999\n",
      "94680 0.43828500000000004 0.970749\n",
      "94714 0.43828500000000004 0.960541\n",
      "94823 0.43828500000000004 0.968073\n",
      "94929 0.43828500000000004 0.991667\n",
      "94949 0.43828500000000004 0.954936\n",
      "94967 0.43828500000000004 0.952897\n",
      "94968 0.43828500000000004 0.999095\n",
      "95086 0.43828500000000004 0.989247\n",
      "95090 0.43828500000000004 0.972714\n",
      "95462 0.43828500000000004 0.99871\n",
      "95470 0.43828500000000004 0.982851\n",
      "95503 0.43828500000000004 0.970749\n",
      "95519 0.43828500000000004 0.95519\n",
      "95546 0.43828500000000004 0.973566\n",
      "95702 0.43828500000000004 0.983584\n",
      "95715 0.43828500000000004 0.999021\n",
      "95727 0.43828500000000004 0.960429\n",
      "95730 0.43828500000000004 0.95851\n",
      "95787 0.43828500000000004 0.997339\n",
      "95800 0.43828500000000004 0.997909\n",
      "95804 0.43828500000000004 0.994421\n",
      "95814 0.43828500000000004 0.993527\n",
      "95815 0.43828500000000004 0.996613\n",
      "95860 0.43828500000000004 0.970917\n",
      "95894 0.43828500000000004 0.952897\n",
      "96166 0.43828500000000004 0.970917\n",
      "96210 0.43828500000000004 0.964304\n",
      "96308 0.43828500000000004 0.998283\n",
      "96312 0.43828500000000004 0.962637\n",
      "96442 0.43828500000000004 0.95519\n",
      "96511 0.43828500000000004 0.983962\n",
      "96628 0.43828500000000004 0.994421\n",
      "96636 0.43828500000000004 0.966972\n",
      "96857 0.43828500000000004 0.994678\n",
      "96874 0.43828500000000004 0.994545\n",
      "96896 0.43828500000000004 0.992293\n",
      "96919 0.43828500000000004 0.95559\n",
      "96978 0.43828500000000004 0.999432\n",
      "97094 0.43828500000000004 0.970749\n",
      "97144 0.43828500000000004 0.978156\n",
      "97458 0.43828500000000004 0.975227\n",
      "97705 0.43828500000000004 0.998825\n",
      "97713 0.43828500000000004 0.95483\n",
      "97807 0.43828500000000004 0.970917\n",
      "97828 0.43828500000000004 0.961135\n",
      "97887 0.43828500000000004 0.991706\n",
      "97971 0.43828500000000004 0.985735\n",
      "98046 0.43828500000000004 0.995316\n",
      "98205 0.43828500000000004 0.996613\n",
      "98310 0.43828500000000004 0.992289\n",
      "98320 0.43828500000000004 0.993938\n",
      "98340 0.43828500000000004 0.991349\n",
      "98600 0.43828500000000004 0.988928\n",
      "98747 0.43828500000000004 0.994545\n",
      "98802 0.43828500000000004 0.970764\n",
      "98819 0.43828500000000004 0.960429\n",
      "98878 0.43828500000000004 0.982293\n",
      "98994 0.43828500000000004 0.992676\n",
      "99128 0.43828500000000004 0.962637\n",
      "99177 0.43828500000000004 0.952897\n",
      "99221 0.43828500000000004 0.970693\n",
      "99242 0.43828500000000004 0.999002\n",
      "99281 0.43828500000000004 0.972119\n",
      "99288 0.43828500000000004 0.984745\n",
      "99431 0.43828500000000004 0.954936\n",
      "99519 0.43828500000000004 0.993009\n",
      "99626 0.43828500000000004 0.96192\n",
      "99642 0.43828500000000004 0.989409\n",
      "99780 0.43828500000000004 0.999568\n",
      "99781 0.43828500000000004 0.962726\n",
      "99859 0.43828500000000004 0.967767\n",
      "99969 0.43828500000000004 0.991562\n",
      "100066 0.43828500000000004 0.994229\n",
      "100119 0.43828500000000004 0.993837\n",
      "100124 0.43828500000000004 0.997553\n",
      "100147 0.43828500000000004 0.998283\n",
      "100220 0.43828500000000004 0.989659\n",
      "100240 0.43828500000000004 0.997909\n",
      "100248 0.43828500000000004 0.982302\n",
      "100373 0.43828500000000004 0.999095\n",
      "100381 0.43828500000000004 0.989424\n",
      "100404 0.43828500000000004 0.957937\n",
      "100477 0.43828500000000004 0.962726\n",
      "100619 0.43828500000000004 0.959423\n",
      "100714 0.43828500000000004 0.999765\n",
      "100716 0.43828500000000004 0.982302\n",
      "100812 0.43828500000000004 0.981753\n",
      "100921 0.43828500000000004 0.96192\n",
      "101050 0.43828500000000004 0.992661\n",
      "101055 0.43828500000000004 0.975944\n",
      "101066 0.43828500000000004 0.985762\n",
      "101130 0.43828500000000004 0.97162\n",
      "101376 0.43828500000000004 0.98498\n",
      "101451 0.43828500000000004 0.95559\n",
      "101506 0.43828500000000004 0.999783\n",
      "101529 0.43828500000000004 0.987583\n",
      "101533 0.43828500000000004 0.999999\n",
      "101668 0.43828500000000004 0.983464\n",
      "101675 0.43828500000000004 0.991349\n",
      "101689 0.43828500000000004 0.989973\n",
      "101712 0.43828500000000004 0.969403\n",
      "101809 0.43828500000000004 0.972119\n",
      "101860 0.43828500000000004 0.992198\n",
      "101887 0.43828500000000004 0.985728\n",
      "102245 0.43828500000000004 0.994883\n",
      "102299 0.43828500000000004 0.979693\n",
      "102327 0.43828500000000004 0.99433\n",
      "102440 0.43828500000000004 0.970749\n",
      "102476 0.43828500000000004 0.999259\n",
      "102576 0.43828500000000004 0.981646\n",
      "102580 0.43828500000000004 0.976652\n",
      "102938 0.43828500000000004 0.986386\n",
      "102939 0.43828500000000004 0.999095\n",
      "102993 0.43828500000000004 0.992198\n",
      "103095 0.43828500000000004 0.989067\n",
      "103122 0.43828500000000004 0.973566\n",
      "103265 0.43828500000000004 0.992198\n",
      "103305 0.43828500000000004 0.998444\n",
      "103326 0.43828500000000004 0.997909\n",
      "103504 0.43828500000000004 0.956727\n",
      "103673 0.43828500000000004 0.959017\n",
      "103703 0.43828500000000004 0.991554\n",
      "103725 0.43828500000000004 0.997306\n",
      "103827 0.43828500000000004 0.993063\n",
      "103996 0.43828500000000004 0.965786\n",
      "104040 0.43828500000000004 0.966972\n",
      "104163 0.43828500000000004 0.997553\n",
      "104238 0.43828500000000004 0.991554\n",
      "104291 0.43828500000000004 0.991985\n",
      "104301 0.43828500000000004 0.973566\n",
      "104336 0.43828500000000004 0.972119\n",
      "104347 0.43828500000000004 0.997909\n",
      "104389 0.43828500000000004 0.970749\n",
      "104498 0.43828500000000004 0.996613\n",
      "104693 0.43828500000000004 0.987184\n",
      "104722 0.43828500000000004 0.95559\n",
      "104732 0.43828500000000004 0.969998\n",
      "104914 0.43828500000000004 0.987184\n",
      "105113 0.43828500000000004 0.968271\n",
      "105368 0.43828500000000004 0.992786\n",
      "105476 0.43828500000000004 0.997999\n",
      "105624 0.43828500000000004 0.988543\n",
      "105654 0.43828500000000004 0.998796\n",
      "105663 0.43828500000000004 0.965359\n",
      "105667 0.43828500000000004 0.998825\n",
      "105697 0.43828500000000004 0.953228\n",
      "105732 0.43828500000000004 0.963687\n",
      "105759 0.43828500000000004 0.972119\n",
      "105790 0.43828500000000004 0.973944\n",
      "105801 0.43828500000000004 0.991866\n",
      "105853 0.43828500000000004 0.989659\n",
      "105860 0.43828500000000004 0.986803\n",
      "105964 0.43828500000000004 0.985735\n",
      "106040 0.43828500000000004 0.985735\n",
      "106137 0.43828500000000004 0.999708\n",
      "106195 0.43828500000000004 0.988543\n",
      "106308 0.43828500000000004 0.998444\n",
      "106395 0.43828500000000004 0.957033\n",
      "106404 0.43828500000000004 0.962191\n",
      "106423 0.43828500000000004 0.993724\n",
      "106510 0.43828500000000004 0.96192\n",
      "106547 0.43828500000000004 0.967843\n",
      "106639 0.43828500000000004 0.991667\n",
      "106732 0.43828500000000004 0.970749\n",
      "107028 0.43828500000000004 0.99837\n",
      "107061 0.43828500000000004 0.999617\n",
      "107222 0.43828500000000004 0.970749\n",
      "107277 0.43828500000000004 0.991349\n",
      "107366 0.43828500000000004 0.990343\n",
      "107455 0.43828500000000004 0.95519\n",
      "107479 0.43828500000000004 0.993983\n",
      "107508 0.43828500000000004 0.992786\n",
      "107530 0.43828500000000004 0.998283\n",
      "107550 0.43828500000000004 0.981023\n",
      "107638 0.43828500000000004 0.972596\n",
      "107686 0.43828500000000004 0.972119\n",
      "107729 0.43828500000000004 0.992804\n",
      "107860 0.43828500000000004 0.972119\n",
      "107875 0.43828500000000004 0.970369\n",
      "107904 0.43828500000000004 0.972119\n",
      "107964 0.43828500000000004 0.965786\n",
      "108003 0.43828500000000004 0.985762\n",
      "108013 0.43828500000000004 0.987799\n",
      "108100 0.43828500000000004 0.986771\n",
      "108419 0.43828500000000004 0.970917\n",
      "108451 0.43828500000000004 0.998283\n",
      "108512 0.43828500000000004 0.967843\n",
      "108540 0.43828500000000004 0.987184\n",
      "108601 0.43828500000000004 0.999999\n",
      "108611 0.43828500000000004 0.993527\n",
      "108624 0.43828500000000004 0.966972\n",
      "108661 0.43828500000000004 0.970749\n",
      "108723 0.43828500000000004 0.978156\n",
      "108824 0.43828500000000004 0.993724\n",
      "108876 0.43828500000000004 0.993009\n",
      "108918 0.43828500000000004 0.998283\n",
      "108972 0.43828500000000004 0.968467\n",
      "109084 0.43828500000000004 0.95483\n",
      "109170 0.43828500000000004 0.996613\n",
      "109202 0.43828500000000004 0.962637\n",
      "109214 0.43828500000000004 0.988543\n",
      "109344 0.43828500000000004 0.976002\n",
      "109416 0.43828500000000004 0.95559\n",
      "109500 0.43828500000000004 0.966972\n",
      "109580 0.43828500000000004 0.970749\n",
      "109693 0.43828500000000004 0.993508\n",
      "109761 0.43828500000000004 0.95519\n",
      "109810 0.43828500000000004 0.981946\n",
      "109835 0.43828500000000004 0.998525\n",
      "109908 0.43828500000000004 0.957937\n",
      "109921 0.43828500000000004 0.973705\n",
      "109949 0.43828500000000004 0.95519\n",
      "109961 0.43828500000000004 0.960071\n",
      "109967 0.43828500000000004 0.95483\n",
      "110065 0.43828500000000004 0.999573\n",
      "110077 0.43828500000000004 0.970749\n",
      "110340 0.43828500000000004 0.980642\n",
      "110480 0.43828500000000004 0.99792\n",
      "110512 0.43828500000000004 0.990296\n",
      "110561 0.43828500000000004 0.975108\n",
      "110584 0.43828500000000004 0.996613\n",
      "110609 0.43828500000000004 0.987184\n",
      "110610 0.43828500000000004 0.994097\n",
      "110727 0.43828500000000004 0.99792\n",
      "110738 0.43828500000000004 0.998161\n",
      "110941 0.43828500000000004 0.958725\n",
      "111096 0.43828500000000004 0.993181\n",
      "111133 0.43828500000000004 0.993009\n",
      "111170 0.43828500000000004 0.982302\n",
      "111184 0.43828500000000004 0.997479\n",
      "111258 0.43828500000000004 0.991349\n",
      "111293 0.43828500000000004 0.99837\n",
      "111311 0.43828500000000004 0.993724\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "111486 0.43828500000000004 0.971222\n",
      "111668 0.43828500000000004 0.969344\n",
      "111675 0.43828500000000004 0.957937\n",
      "111683 0.43828500000000004 0.998444\n",
      "111765 0.43828500000000004 0.960541\n",
      "111834 0.43828500000000004 0.99837\n",
      "111863 0.43828500000000004 0.996718\n",
      "111931 0.43828500000000004 0.989948\n",
      "111975 0.43828500000000004 0.974514\n",
      "111999 0.43828500000000004 0.95519\n",
      "112103 0.43828500000000004 0.962637\n",
      "112120 0.43828500000000004 0.997258\n",
      "112129 0.43828500000000004 0.972557\n",
      "112182 0.43828500000000004 0.994097\n",
      "112200 0.43828500000000004 0.99837\n",
      "112277 0.43828500000000004 0.99462\n",
      "112528 0.43828500000000004 0.964304\n",
      "112628 0.43828500000000004 0.964994\n",
      "112685 0.43828500000000004 0.992531\n",
      "112689 0.43828500000000004 0.982302\n",
      "112783 0.43828500000000004 0.962637\n",
      "112819 0.43828500000000004 0.989659\n",
      "112838 0.43828500000000004 0.962637\n",
      "112907 0.43828500000000004 0.983464\n",
      "113022 0.43828500000000004 0.981023\n",
      "113072 0.43828500000000004 0.95519\n",
      "113176 0.43828500000000004 1.0\n",
      "113191 0.43828500000000004 0.971264\n",
      "113283 0.43828500000000004 0.962637\n",
      "113294 0.43828500000000004 0.979956\n",
      "113376 0.43828500000000004 0.98473\n",
      "113497 0.43828500000000004 0.970749\n",
      "113507 0.43828500000000004 0.95559\n",
      "113575 0.43828500000000004 0.956727\n",
      "113602 0.43828500000000004 0.973566\n",
      "113778 0.43828500000000004 0.990343\n",
      "113835 0.43828500000000004 0.973566\n",
      "113848 0.43828500000000004 0.959748\n",
      "113979 0.43828500000000004 0.962637\n",
      "114147 0.43828500000000004 0.962726\n",
      "114159 0.43828500000000004 0.985574\n",
      "114185 0.43828500000000004 0.95483\n",
      "114190 0.43828500000000004 0.989659\n",
      "114196 0.43828500000000004 0.991667\n",
      "114238 0.43828500000000004 0.974514\n",
      "114293 0.43828500000000004 0.978366\n",
      "114407 0.43828500000000004 0.991866\n",
      "114545 0.43828500000000004 0.994883\n",
      "114590 0.43828500000000004 0.970749\n",
      "114622 0.43828500000000004 0.954531\n",
      "114650 0.43828500000000004 0.966175\n",
      "114656 0.43828500000000004 0.972398\n",
      "114681 0.43828500000000004 0.960071\n",
      "114704 0.43828500000000004 0.960354\n",
      "114736 0.43828500000000004 0.973604\n",
      "114763 0.43828500000000004 0.982293\n",
      "114790 0.43828500000000004 0.966972\n",
      "114826 0.43828500000000004 0.993009\n",
      "114830 0.43828500000000004 0.972596\n",
      "114904 0.43828500000000004 0.972119\n",
      "114924 0.43828500000000004 0.973001\n",
      "115106 0.43828500000000004 0.966972\n",
      "115133 0.43828500000000004 0.987559\n",
      "115146 0.43828500000000004 0.987184\n",
      "115206 0.43828500000000004 0.999987\n",
      "115388 0.43828500000000004 0.994994\n",
      "115469 0.43828500000000004 0.959423\n",
      "115619 0.43828500000000004 0.965359\n",
      "115688 0.43828500000000004 0.961557\n",
      "115715 0.43828500000000004 0.965786\n",
      "115728 0.43828500000000004 0.953775\n",
      "115933 0.43828500000000004 0.989659\n",
      "115979 0.43828500000000004 0.95519\n",
      "116012 0.43828500000000004 0.994888\n",
      "116018 0.43828500000000004 0.962637\n",
      "116077 0.43828500000000004 0.996914\n",
      "116097 0.43828500000000004 0.960335\n",
      "116378 0.43828500000000004 0.957937\n",
      "116417 0.43828500000000004 0.990343\n",
      "116440 0.43828500000000004 0.997553\n",
      "116573 0.43828500000000004 0.970369\n",
      "116608 0.43828500000000004 0.960541\n",
      "116690 0.43828500000000004 0.962726\n",
      "116994 0.43828500000000004 0.999095\n",
      "117044 0.43828500000000004 0.965011\n",
      "117053 0.43828500000000004 0.9817\n",
      "117095 0.43828500000000004 0.99858\n",
      "117119 0.43828500000000004 0.984362\n",
      "117316 0.43828500000000004 0.960071\n",
      "117320 0.43828500000000004 0.997553\n",
      "117336 0.43828500000000004 0.961557\n",
      "117537 0.43828500000000004 0.962637\n",
      "117541 0.43828500000000004 0.997098\n",
      "117585 0.43828500000000004 0.991866\n",
      "117692 0.43828500000000004 0.991866\n",
      "117712 0.43828500000000004 0.965359\n",
      "117758 0.43828500000000004 0.971944\n",
      "117778 0.43828500000000004 0.960354\n",
      "117812 0.43828500000000004 0.964338\n",
      "117844 0.43828500000000004 0.973944\n",
      "118107 0.43828500000000004 0.970369\n",
      "118121 0.43828500000000004 0.960541\n",
      "118145 0.43828500000000004 0.985969\n",
      "118157 0.43828500000000004 0.999021\n",
      "118158 0.43828500000000004 0.952897\n",
      "118173 0.43828500000000004 0.997479\n",
      "118280 0.43828500000000004 0.967843\n",
      "118344 0.43828500000000004 0.964304\n",
      "118399 0.43828500000000004 0.959423\n",
      "118506 0.43828500000000004 0.99837\n",
      "118513 0.43828500000000004 0.980539\n",
      "118544 0.43828500000000004 0.997909\n",
      "118568 0.43828500000000004 0.990343\n",
      "118569 0.43828500000000004 0.973566\n",
      "118666 0.43828500000000004 0.999827\n",
      "118680 0.43828500000000004 0.981753\n",
      "118778 0.43828500000000004 0.978314\n",
      "118779 0.43828500000000004 0.970917\n",
      "118788 0.43828500000000004 0.990343\n",
      "119236 0.43828500000000004 0.993938\n",
      "119255 0.43828500000000004 0.976002\n",
      "119282 0.43828500000000004 0.988029\n",
      "119329 0.43828500000000004 0.961135\n",
      "119337 0.43828500000000004 0.999708\n",
      "119355 0.43828500000000004 0.996334\n",
      "119578 0.43828500000000004 0.957937\n",
      "119677 0.43828500000000004 0.987184\n",
      "119689 0.43828500000000004 0.975227\n",
      "119702 0.43828500000000004 0.953448\n",
      "119755 0.43828500000000004 0.967767\n",
      "119874 0.43828500000000004 0.972119\n",
      "119894 0.43828500000000004 0.99805\n",
      "119924 0.43828500000000004 0.991349\n",
      "119978 0.43828500000000004 0.970764\n",
      "119988 0.43828500000000004 0.988151\n",
      "120006 0.43828500000000004 0.954509\n",
      "120048 0.43828500000000004 0.968073\n",
      "120080 0.43828500000000004 0.992289\n",
      "120144 0.43828500000000004 0.960429\n",
      "120153 0.43828500000000004 0.997734\n",
      "120259 0.43828500000000004 0.998444\n",
      "120290 0.43828500000000004 0.970749\n",
      "120338 0.43828500000000004 0.995723\n",
      "120426 0.43828500000000004 0.967997\n",
      "120440 0.43828500000000004 0.966972\n",
      "120516 0.43828500000000004 0.973566\n",
      "120550 0.43828500000000004 0.997761\n",
      "120676 0.43828500000000004 0.988543\n",
      "120749 0.43828500000000004 0.981753\n",
      "120766 0.43828500000000004 0.963078\n",
      "120773 0.43828500000000004 0.999971\n",
      "120816 0.43828500000000004 0.991667\n",
      "120883 0.43828500000000004 0.95519\n",
      "121242 0.43828500000000004 0.99837\n",
      "121320 0.43828500000000004 0.965786\n",
      "121337 0.43828500000000004 0.980111\n",
      "121354 0.43828500000000004 0.970917\n",
      "121402 0.43828500000000004 0.95519\n",
      "121443 0.43828500000000004 0.997339\n",
      "121512 0.43828500000000004 0.969344\n",
      "121561 0.43828500000000004 0.991349\n",
      "121573 0.43828500000000004 0.970749\n",
      "121596 0.43828500000000004 0.970369\n",
      "121659 0.43828500000000004 0.970693\n",
      "121736 0.43828500000000004 0.973604\n",
      "121854 0.43828500000000004 0.993343\n",
      "121871 0.43828500000000004 0.997909\n",
      "121956 0.43828500000000004 0.99871\n",
      "121991 0.43828500000000004 0.994883\n",
      "122029 0.43828500000000004 0.982302\n",
      "122039 0.43828500000000004 0.991706\n",
      "122127 0.43828500000000004 0.95559\n",
      "122201 0.43828500000000004 0.997479\n",
      "122230 0.43828500000000004 0.993938\n",
      "122234 0.43828500000000004 0.95519\n",
      "122244 0.43828500000000004 0.959423\n",
      "122256 0.43828500000000004 0.980111\n",
      "122334 0.43828500000000004 0.986152\n",
      "122355 0.43828500000000004 0.993009\n",
      "122378 0.43828500000000004 0.998257\n",
      "122385 0.43828500000000004 0.99792\n",
      "122551 0.43828500000000004 0.993009\n",
      "122566 0.43828500000000004 0.987184\n",
      "122621 0.43828500000000004 0.979693\n",
      "122716 0.43828500000000004 0.998257\n",
      "122801 0.43828500000000004 0.952897\n",
      "122878 0.43828500000000004 0.979276\n",
      "122882 0.43828500000000004 0.987184\n",
      "122962 0.43828500000000004 0.961557\n",
      "123024 0.43828500000000004 0.963078\n",
      "123076 0.43828500000000004 0.965011\n",
      "123184 0.43828500000000004 0.95851\n",
      "123319 0.43828500000000004 0.970749\n",
      "123332 0.43828500000000004 0.997693\n",
      "123357 0.43828500000000004 0.99462\n",
      "123394 0.43828500000000004 0.991667\n",
      "123495 0.43828500000000004 0.954936\n",
      "123514 0.43828500000000004 0.990853\n",
      "123523 0.43828500000000004 0.982302\n",
      "123553 0.43828500000000004 0.997027\n",
      "123557 0.43828500000000004 0.967767\n",
      "123568 0.43828500000000004 0.965011\n",
      "123603 0.43828500000000004 0.98981\n",
      "123630 0.43828500000000004 0.997553\n",
      "123704 0.43828500000000004 0.960541\n",
      "123742 0.43828500000000004 0.990343\n",
      "123752 0.43828500000000004 0.991886\n",
      "123773 0.43828500000000004 0.972119\n",
      "123946 0.43828500000000004 0.991349\n",
      "124205 0.43828500000000004 0.99964\n",
      "124216 0.43828500000000004 0.985735\n",
      "124287 0.43828500000000004 0.96192\n",
      "124364 0.43828500000000004 0.985969\n",
      "124472 0.43828500000000004 0.985762\n",
      "124525 0.43828500000000004 0.998101\n",
      "124588 0.43828500000000004 0.990343\n",
      "124595 0.43828500000000004 0.977482\n",
      "124643 0.43828500000000004 0.979693\n",
      "124753 0.43828500000000004 0.960541\n",
      "124812 0.43828500000000004 0.993297\n",
      "124895 0.43828500000000004 0.968073\n",
      "124953 0.43828500000000004 0.997999\n",
      "125028 0.43828500000000004 0.996931\n",
      "125197 0.43828500000000004 0.970917\n",
      "125292 0.43828500000000004 0.961283\n",
      "125313 0.43828500000000004 0.970917\n",
      "125422 0.43828500000000004 0.980642\n",
      "125463 0.43828500000000004 0.965359\n",
      "125548 0.43828500000000004 0.973566\n",
      "125578 0.43828500000000004 0.985762\n",
      "125599 0.43828500000000004 0.999953\n",
      "125626 0.43828500000000004 0.996094\n",
      "125668 0.43828500000000004 0.999996\n",
      "125671 0.43828500000000004 0.95483\n",
      "125828 0.43828500000000004 0.980829\n",
      "125913 0.43828500000000004 0.982851\n",
      "125918 0.43828500000000004 0.993938\n",
      "125952 0.43828500000000004 0.993009\n",
      "126107 0.43828500000000004 0.972557\n",
      "126407 0.43828500000000004 0.999949\n",
      "126409 0.43828500000000004 0.991579\n",
      "126415 0.43828500000000004 0.957033\n",
      "126623 0.43828500000000004 0.998525\n",
      "126664 0.43828500000000004 0.982302\n",
      "126717 0.43828500000000004 0.996309\n",
      "126809 0.43828500000000004 0.97773\n",
      "126836 0.43828500000000004 0.962637\n",
      "127004 0.43828500000000004 0.990343\n",
      "127189 0.43828500000000004 0.954531\n",
      "127311 0.43828500000000004 0.997553\n",
      "127341 0.43828500000000004 0.987516\n",
      "127560 0.43828500000000004 0.972596\n",
      "127772 0.43828500000000004 0.995843\n",
      "127853 0.43828500000000004 0.966909\n",
      "127932 0.43828500000000004 0.953622\n",
      "128157 0.43828500000000004 0.982302\n",
      "128167 0.43828500000000004 0.973489\n",
      "128254 0.43828500000000004 0.997909\n",
      "128349 0.43828500000000004 0.956727\n",
      "128440 0.43828500000000004 0.994883\n",
      "128514 0.43828500000000004 0.996094\n",
      "128516 0.43828500000000004 1.0\n",
      "128551 0.43828500000000004 0.997553\n",
      "128560 0.43828500000000004 0.999147\n",
      "128563 0.43828500000000004 0.988543\n",
      "128571 0.43828500000000004 0.982302\n",
      "128601 0.43828500000000004 0.983131\n",
      "128728 0.43828500000000004 0.997098\n",
      "128780 0.43828500000000004 0.992266\n",
      "128805 0.43828500000000004 0.99837\n",
      "128881 0.43828500000000004 0.991866\n",
      "128882 0.43828500000000004 0.993009\n",
      "128897 0.43828500000000004 0.970764\n",
      "128971 0.43828500000000004 0.999115\n",
      "129149 0.43828500000000004 0.952897\n",
      "129172 0.43828500000000004 0.991562\n",
      "129176 0.43828500000000004 0.987583\n",
      "129182 0.43828500000000004 0.985574\n",
      "129193 0.43828500000000004 0.986583\n",
      "129203 0.43828500000000004 0.979693\n",
      "129205 0.43828500000000004 0.98105\n",
      "129214 0.43828500000000004 0.967843\n",
      "129418 0.43828500000000004 0.966972\n",
      "129478 0.43828500000000004 0.952897\n",
      "129482 0.43828500000000004 0.995754\n",
      "129568 0.43828500000000004 0.994888\n",
      "129707 0.43828500000000004 0.987583\n",
      "129800 0.43828500000000004 0.986803\n",
      "129802 0.43828500000000004 0.971222\n",
      "129810 0.43828500000000004 0.953228\n",
      "129824 0.43828500000000004 0.965359\n",
      "129912 0.43828500000000004 0.990026\n",
      "130114 0.43828500000000004 0.979276\n",
      "130148 0.43828500000000004 0.992676\n",
      "130187 0.43828500000000004 0.992439\n",
      "130244 0.43828500000000004 0.965622\n",
      "130527 0.43828500000000004 0.972596\n",
      "130546 0.43828500000000004 0.998307\n",
      "130580 0.43828500000000004 0.979693\n",
      "130702 0.43828500000000004 0.980359\n",
      "130847 0.43828500000000004 0.95559\n",
      "130903 0.43828500000000004 0.989659\n",
      "130955 0.43828500000000004 0.959423\n",
      "131027 0.43828500000000004 0.973604\n",
      "131080 0.43828500000000004 0.977514\n",
      "131127 0.43828500000000004 0.994532\n",
      "131193 0.43828500000000004 0.999259\n",
      "131232 0.43828500000000004 0.998684\n",
      "131242 0.43828500000000004 0.965786\n",
      "131253 0.43828500000000004 0.982302\n",
      "131368 0.43828500000000004 0.981023\n",
      "131408 0.43828500000000004 0.998684\n",
      "131449 0.43828500000000004 0.985504\n",
      "131457 0.43828500000000004 0.965359\n",
      "131480 0.43828500000000004 0.988414\n",
      "131491 0.43828500000000004 0.989247\n",
      "131624 0.43828500000000004 0.972119\n",
      "131648 0.43828500000000004 0.993837\n",
      "131777 0.43828500000000004 0.978574\n",
      "131826 0.43828500000000004 0.987662\n",
      "131882 0.43828500000000004 0.984811\n",
      "131931 0.43828500000000004 0.973566\n",
      "131964 0.43828500000000004 0.972119\n",
      "132021 0.43828500000000004 0.997429\n",
      "132202 0.43828500000000004 0.966972\n",
      "132422 0.43828500000000004 0.997999\n",
      "132497 0.43828500000000004 0.990026\n",
      "132750 0.43828500000000004 0.990853\n",
      "132865 0.43828500000000004 0.984811\n",
      "132896 0.43828500000000004 0.989409\n",
      "132915 0.43828500000000004 0.967843\n",
      "132980 0.43828500000000004 0.95519\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "133026 0.43828500000000004 0.969344\n",
      "133202 0.43828500000000004 0.98473\n",
      "133242 0.43828500000000004 0.994421\n",
      "133264 0.43828500000000004 0.997847\n",
      "133295 0.43828500000000004 0.95519\n",
      "133324 0.43828500000000004 0.95519\n",
      "133527 0.43828500000000004 0.992804\n",
      "133537 0.43828500000000004 0.970369\n",
      "133636 0.43828500000000004 0.997734\n",
      "133665 0.43828500000000004 0.991579\n",
      "133736 0.43828500000000004 0.95519\n",
      "133797 0.43828500000000004 0.991302\n",
      "133840 0.43828500000000004 0.984377\n",
      "133846 0.43828500000000004 0.994883\n",
      "133886 0.43828500000000004 0.965011\n",
      "133974 0.43828500000000004 0.999021\n",
      "134064 0.43828500000000004 0.997693\n",
      "134118 0.43828500000000004 0.991349\n",
      "134151 0.43828500000000004 0.997339\n",
      "134264 0.43828500000000004 0.983584\n",
      "134307 0.43828500000000004 0.960071\n",
      "134325 0.43828500000000004 0.95519\n",
      "134355 0.43828500000000004 0.965359\n",
      "134422 0.43828500000000004 0.978103\n",
      "134435 0.43828500000000004 0.997553\n",
      "134436 0.43828500000000004 0.985269\n",
      "134462 0.43828500000000004 0.988928\n",
      "134473 0.43828500000000004 0.989735\n",
      "134496 0.43828500000000004 0.965786\n",
      "134548 0.43828500000000004 0.987184\n",
      "134563 0.43828500000000004 0.989659\n",
      "134647 0.43828500000000004 0.973566\n",
      "134648 0.43828500000000004 1.0\n",
      "134818 0.43828500000000004 0.997299\n",
      "134838 0.43828500000000004 0.965011\n",
      "134841 0.43828500000000004 0.973705\n",
      "134847 0.43828500000000004 0.998784\n",
      "134876 0.43828500000000004 0.965425\n",
      "135024 0.43828500000000004 0.982302\n",
      "135106 0.43828500000000004 0.957033\n",
      "135192 0.43828500000000004 0.972119\n",
      "135193 0.43828500000000004 0.95519\n",
      "135214 0.43828500000000004 0.991902\n",
      "135220 0.43828500000000004 0.970749\n",
      "135316 0.43828500000000004 0.993527\n",
      "135321 0.43828500000000004 0.98818\n",
      "135330 0.43828500000000004 0.970917\n",
      "135597 0.43828500000000004 0.952897\n",
      "135814 0.43828500000000004 0.993938\n",
      "135945 0.43828500000000004 0.957937\n",
      "135968 0.43828500000000004 0.966972\n",
      "135999 0.43828500000000004 0.993009\n",
      "136065 0.43828500000000004 0.982302\n",
      "136090 0.43828500000000004 0.991681\n",
      "136295 0.43828500000000004 1.0\n",
      "136371 0.43828500000000004 0.997909\n",
      "136444 0.43828500000000004 0.999021\n",
      "136470 0.43828500000000004 0.967843\n",
      "136536 0.43828500000000004 0.970917\n",
      "136618 0.43828500000000004 0.985454\n",
      "136720 0.43828500000000004 0.973944\n",
      "136866 0.43828500000000004 0.998671\n",
      "136945 0.43828500000000004 0.992439\n",
      "137039 0.43828500000000004 0.953884\n",
      "137282 0.43828500000000004 0.960429\n",
      "137290 0.43828500000000004 0.998444\n",
      "137328 0.43828500000000004 0.997504\n",
      "137416 0.43828500000000004 0.981806\n",
      "137467 0.43828500000000004 0.989659\n",
      "137619 0.43828500000000004 0.954936\n",
      "137727 0.43828500000000004 0.997553\n",
      "137855 0.43828500000000004 0.990343\n",
      "137865 0.43828500000000004 0.994229\n",
      "137898 0.43828500000000004 0.990736\n",
      "137950 0.43828500000000004 0.997999\n",
      "137998 0.43828500000000004 0.997504\n",
      "138025 0.43828500000000004 0.95483\n",
      "138038 0.43828500000000004 0.969197\n",
      "138046 0.43828500000000004 0.980539\n",
      "138122 0.43828500000000004 0.977482\n",
      "138484 0.43828500000000004 0.982302\n",
      "138500 0.43828500000000004 0.996307\n",
      "138564 0.43828500000000004 0.960541\n",
      "138598 0.43828500000000004 0.967843\n",
      "138626 0.43828500000000004 0.960579\n",
      "138644 0.43828500000000004 0.95519\n",
      "138675 0.43828500000000004 0.954316\n",
      "138756 0.43828500000000004 0.980539\n",
      "138783 0.43828500000000004 0.982293\n",
      "138789 0.43828500000000004 0.965786\n",
      "138818 0.43828500000000004 0.975653\n",
      "138852 0.43828500000000004 0.99871\n",
      "138868 0.43828500000000004 0.965786\n",
      "138890 0.43828500000000004 0.993539\n",
      "139016 0.43828500000000004 0.989659\n",
      "139017 0.43828500000000004 0.981753\n",
      "139018 0.43828500000000004 0.980958\n",
      "139173 0.43828500000000004 0.982302\n",
      "139211 0.43828500000000004 0.989659\n",
      "139327 0.43828500000000004 0.988543\n",
      "139331 0.43828500000000004 0.998283\n",
      "139452 0.43828500000000004 0.965359\n",
      "139685 0.43828500000000004 0.998684\n",
      "139690 0.43828500000000004 0.989948\n",
      "140094 0.43828500000000004 0.973566\n",
      "140177 0.43828500000000004 0.954531\n",
      "140384 0.43828500000000004 0.999021\n",
      "140399 0.43828500000000004 0.965044\n",
      "140421 0.43828500000000004 0.952897\n",
      "140494 0.43828500000000004 0.970369\n",
      "140518 0.43828500000000004 0.973001\n",
      "140524 0.43828500000000004 0.988841\n",
      "140636 0.43828500000000004 0.980829\n",
      "140645 0.43828500000000004 0.96192\n",
      "140678 0.43828500000000004 0.96192\n",
      "140741 0.43828500000000004 0.973705\n",
      "140786 0.43828500000000004 0.991349\n",
      "140820 0.43828500000000004 0.996808\n",
      "140855 0.43828500000000004 0.960541\n",
      "140905 0.43828500000000004 0.973566\n",
      "141001 0.43828500000000004 0.987184\n",
      "141092 0.43828500000000004 0.998146\n",
      "141304 0.43828500000000004 0.997384\n",
      "141334 0.43828500000000004 0.999765\n",
      "141389 0.43828500000000004 0.987516\n",
      "141509 0.43828500000000004 0.962637\n",
      "141605 0.43828500000000004 0.952897\n",
      "141684 0.43828500000000004 0.973705\n",
      "141801 0.43828500000000004 0.991886\n",
      "142032 0.43828500000000004 0.989003\n",
      "142101 0.43828500000000004 0.992676\n",
      "142116 0.43828500000000004 0.992786\n",
      "142146 0.43828500000000004 0.997429\n",
      "142307 0.43828500000000004 0.973566\n",
      "142506 0.43828500000000004 0.95483\n",
      "142715 0.43828500000000004 0.964338\n",
      "142975 0.43828500000000004 0.989659\n",
      "142988 0.43828500000000004 0.999002\n",
      "143119 0.43828500000000004 0.979693\n",
      "143144 0.43828500000000004 0.960071\n",
      "143209 0.43828500000000004 0.997479\n",
      "143318 0.43828500000000004 0.988543\n",
      "143332 0.43828500000000004 0.988543\n",
      "143366 0.43828500000000004 0.989659\n",
      "143475 0.43828500000000004 0.96521\n",
      "143511 0.43828500000000004 0.987799\n",
      "143515 0.43828500000000004 0.980642\n",
      "143560 0.43828500000000004 0.970917\n",
      "143599 0.43828500000000004 0.998684\n",
      "143607 0.43828500000000004 0.983464\n",
      "143635 0.43828500000000004 0.952897\n",
      "143694 0.43828500000000004 0.979693\n",
      "143702 0.43828500000000004 0.980642\n",
      "143713 0.43828500000000004 0.992676\n",
      "143717 0.43828500000000004 0.962637\n",
      "143796 0.43828500000000004 0.991667\n",
      "143838 0.43828500000000004 0.960071\n",
      "143841 0.43828500000000004 0.957937\n",
      "143859 0.43828500000000004 0.984877\n",
      "143861 0.43828500000000004 0.980642\n",
      "143869 0.43828500000000004 0.988145\n",
      "143938 0.43828500000000004 0.966175\n",
      "143962 0.43828500000000004 1.0\n",
      "143982 0.43828500000000004 0.954936\n",
      "144017 0.43828500000000004 0.989948\n",
      "144105 0.43828500000000004 0.990853\n",
      "144168 0.43828500000000004 0.998684\n",
      "144199 0.43828500000000004 0.988543\n",
      "144244 0.43828500000000004 0.980359\n",
      "144296 0.43828500000000004 0.970693\n",
      "144301 0.43828500000000004 0.965786\n",
      "144389 0.43828500000000004 0.987184\n",
      "144408 0.43828500000000004 0.972596\n",
      "144466 0.43828500000000004 0.973066\n",
      "144483 0.43828500000000004 0.990343\n",
      "144492 0.43828500000000004 0.9938\n",
      "144601 0.43828500000000004 0.953884\n",
      "144900 0.43828500000000004 0.953622\n",
      "144916 0.43828500000000004 0.991667\n",
      "145017 0.43828500000000004 0.978324\n",
      "145146 0.43828500000000004 0.954936\n",
      "145149 0.43828500000000004 0.998292\n",
      "145179 0.43828500000000004 0.999891\n",
      "145184 0.43828500000000004 0.997306\n",
      "145484 0.43828500000000004 0.952897\n",
      "145501 0.43828500000000004 0.999259\n",
      "145621 0.43828500000000004 0.956727\n",
      "145698 0.43828500000000004 0.95519\n",
      "145716 0.43828500000000004 0.961135\n",
      "146055 0.43828500000000004 0.993343\n",
      "146095 0.43828500000000004 0.970764\n",
      "146203 0.43828500000000004 0.993508\n",
      "146277 0.43828500000000004 0.988543\n",
      "146287 0.43828500000000004 0.988841\n",
      "146290 0.43828500000000004 0.960071\n",
      "146297 0.43828500000000004 0.999021\n",
      "146300 0.43828500000000004 0.993009\n",
      "146342 0.43828500000000004 0.988543\n",
      "146497 0.43828500000000004 0.997734\n",
      "146560 0.43828500000000004 0.972119\n",
      "146561 0.43828500000000004 0.965011\n",
      "146613 0.43828500000000004 0.998292\n",
      "146639 0.43828500000000004 0.994883\n",
      "146699 0.43828500000000004 0.964338\n",
      "146715 0.43828500000000004 0.964202\n",
      "146723 0.43828500000000004 0.998537\n",
      "146749 0.43828500000000004 0.997693\n",
      "146830 0.43828500000000004 0.95519\n",
      "146919 0.43828500000000004 0.970749\n",
      "146957 0.43828500000000004 0.996914\n",
      "147002 0.43828500000000004 0.979276\n",
      "147016 0.43828500000000004 0.994883\n",
      "147019 0.43828500000000004 0.982302\n",
      "147156 0.43828500000000004 0.970749\n",
      "147175 0.43828500000000004 0.953448\n",
      "147349 0.43828500000000004 0.996094\n",
      "147354 0.43828500000000004 0.989659\n",
      "147536 0.43828500000000004 0.961135\n",
      "147562 0.43828500000000004 0.998105\n",
      "147568 0.43828500000000004 0.996718\n",
      "147698 0.43828500000000004 0.952897\n",
      "147757 0.43828500000000004 0.962191\n",
      "147761 0.43828500000000004 0.970749\n",
      "147762 0.43828500000000004 0.978923\n",
      "147847 0.43828500000000004 0.992446\n",
      "147876 0.43828500000000004 0.970917\n",
      "147921 0.43828500000000004 0.97773\n",
      "147968 0.43828500000000004 0.994883\n",
      "148051 0.43828500000000004 0.982293\n",
      "148086 0.43828500000000004 0.999314\n",
      "148104 0.43828500000000004 0.98473\n",
      "148139 0.43828500000000004 0.962548\n",
      "148160 0.43828500000000004 0.979761\n",
      "148242 0.43828500000000004 0.99965\n",
      "148272 0.43828500000000004 0.990343\n",
      "148302 0.43828500000000004 0.961135\n",
      "148339 0.43828500000000004 0.991349\n",
      "148401 0.43828500000000004 0.976403\n",
      "148496 0.43828500000000004 0.994994\n",
      "148513 0.43828500000000004 0.957738\n",
      "148588 0.43828500000000004 0.9938\n",
      "148699 0.43828500000000004 0.966972\n",
      "148797 0.43828500000000004 0.993938\n",
      "148889 0.43828500000000004 0.983131\n",
      "148932 0.43828500000000004 0.990343\n",
      "148963 0.43828500000000004 0.954316\n",
      "148980 0.43828500000000004 1.0\n",
      "148982 0.43828500000000004 0.998825\n",
      "148991 0.43828500000000004 0.994994\n",
      "149005 0.43828500000000004 0.995762\n",
      "149009 0.43828500000000004 0.970749\n",
      "149046 0.43828500000000004 0.989948\n",
      "149166 0.43828500000000004 0.986771\n",
      "149184 0.43828500000000004 0.984877\n",
      "149197 0.43828500000000004 0.983464\n",
      "149444 0.43828500000000004 0.970369\n",
      "149477 0.43828500000000004 0.991349\n",
      "149603 0.43828500000000004 0.999095\n",
      "149605 0.43828500000000004 0.970764\n",
      "149693 0.43828500000000004 0.95519\n",
      "149793 0.43828500000000004 1.0\n",
      "149881 0.43828500000000004 0.993938\n",
      "150011 0.43828500000000004 0.97518\n",
      "150168 0.43828500000000004 0.957033\n",
      "150327 0.43828500000000004 0.978156\n",
      "150461 0.43828500000000004 0.988841\n",
      "150483 0.43828500000000004 0.994097\n",
      "150491 0.43828500000000004 0.970749\n",
      "150509 0.43828500000000004 0.997553\n",
      "150512 0.43828500000000004 0.979761\n",
      "150641 0.43828500000000004 0.95483\n",
      "150786 0.43828500000000004 0.992439\n",
      "150845 0.43828500000000004 0.991886\n",
      "150909 0.43828500000000004 0.977482\n",
      "150953 0.43828500000000004 0.995843\n",
      "150991 0.43828500000000004 0.988543\n",
      "151125 0.43828500000000004 0.995723\n",
      "151168 0.43828500000000004 0.999624\n",
      "151172 0.43828500000000004 0.97518\n",
      "151188 0.43828500000000004 0.970917\n",
      "151225 0.43828500000000004 0.972119\n",
      "151271 0.43828500000000004 0.999948\n",
      "151310 0.43828500000000004 0.964338\n",
      "151364 0.43828500000000004 0.988841\n",
      "151413 0.43828500000000004 0.954936\n",
      "151433 0.43828500000000004 0.95483\n",
      "151451 0.43828500000000004 0.999147\n",
      "151482 0.43828500000000004 0.975108\n",
      "151561 0.43828500000000004 0.955232\n",
      "151999 0.43828500000000004 0.985735\n",
      "152082 0.43828500000000004 0.989003\n",
      "152115 0.43828500000000004 0.982302\n",
      "152146 0.43828500000000004 0.970917\n",
      "152221 0.43828500000000004 0.973944\n",
      "152296 0.43828500000000004 0.999573\n",
      "152568 0.43828500000000004 0.993724\n",
      "152611 0.43828500000000004 0.989659\n",
      "152830 0.43828500000000004 0.965948\n",
      "152912 0.43828500000000004 0.998283\n",
      "153031 0.43828500000000004 0.993009\n",
      "153050 0.43828500000000004 0.968527\n",
      "153186 0.43828500000000004 0.98473\n",
      "153399 0.43828500000000004 0.972936\n",
      "153550 0.43828500000000004 0.994421\n",
      "153598 0.43828500000000004 0.973066\n",
      "153826 0.43828500000000004 0.998525\n",
      "153829 0.43828500000000004 0.982302\n",
      "153969 0.43828500000000004 0.957937\n",
      "154151 0.43828500000000004 0.95559\n",
      "154152 0.43828500000000004 0.957738\n",
      "154506 0.43828500000000004 0.992676\n",
      "154514 0.43828500000000004 0.967843\n",
      "154543 0.43828500000000004 0.967843\n",
      "154707 0.43828500000000004 0.989659\n",
      "154770 0.43828500000000004 0.994097\n",
      "154834 0.43828500000000004 0.952897\n",
      "154905 0.43828500000000004 0.959423\n",
      "154910 0.43828500000000004 0.970917\n",
      "154943 0.43828500000000004 0.993837\n",
      "154956 0.43828500000000004 0.989948\n",
      "155117 0.43828500000000004 0.960541\n",
      "155149 0.43828500000000004 0.989659\n",
      "155284 0.43828500000000004 0.982302\n",
      "155408 0.43828500000000004 0.979693\n",
      "155427 0.43828500000000004 0.998444\n",
      "155429 0.43828500000000004 0.962637\n",
      "155487 0.43828500000000004 0.972596\n",
      "155785 0.43828500000000004 0.992439\n",
      "155888 0.43828500000000004 0.964117\n",
      "155971 0.43828500000000004 0.970369\n",
      "156007 0.43828500000000004 0.999975\n",
      "156084 0.43828500000000004 0.972119\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "156104 0.43828500000000004 0.989659\n",
      "156260 0.43828500000000004 0.997734\n",
      "156273 0.43828500000000004 0.972596\n",
      "156376 0.43828500000000004 0.95519\n",
      "156401 0.43828500000000004 0.953884\n",
      "156447 0.43828500000000004 0.989424\n",
      "156521 0.43828500000000004 0.95519\n",
      "156595 0.43828500000000004 0.986386\n",
      "156614 0.43828500000000004 0.989003\n",
      "156630 0.43828500000000004 0.970369\n",
      "156791 0.43828500000000004 0.986681\n",
      "156810 0.43828500000000004 0.965011\n",
      "156866 0.43828500000000004 0.998678\n",
      "156919 0.43828500000000004 0.989659\n",
      "156922 0.43828500000000004 0.953622\n",
      "157007 0.43828500000000004 0.994545\n",
      "157052 0.43828500000000004 0.980642\n",
      "157065 0.43828500000000004 0.988029\n",
      "157219 0.43828500000000004 0.992198\n",
      "157237 0.43828500000000004 0.979956\n",
      "157428 0.43828500000000004 0.99837\n",
      "157477 0.43828500000000004 0.980642\n",
      "157629 0.43828500000000004 0.982851\n",
      "157649 0.43828500000000004 0.970917\n",
      "157723 0.43828500000000004 0.985728\n",
      "157754 0.43828500000000004 0.991667\n",
      "157770 0.43828500000000004 0.998101\n",
      "157846 0.43828500000000004 0.988841\n",
      "157853 0.43828500000000004 0.989659\n",
      "157876 0.43828500000000004 0.967767\n",
      "157902 0.43828500000000004 0.967843\n",
      "158077 0.43828500000000004 0.970749\n",
      "158126 0.43828500000000004 0.983679\n",
      "158247 0.43828500000000004 0.975712\n",
      "158259 0.43828500000000004 0.991667\n",
      "158264 0.43828500000000004 0.997553\n",
      "158323 0.43828500000000004 0.979585\n",
      "158356 0.43828500000000004 0.994421\n",
      "158371 0.43828500000000004 0.991866\n",
      "158391 0.43828500000000004 0.98473\n",
      "158411 0.43828500000000004 0.982302\n",
      "158433 0.43828500000000004 0.996155\n",
      "158541 0.43828500000000004 0.95519\n",
      "158548 0.43828500000000004 0.991902\n",
      "158572 0.43828500000000004 0.993953\n",
      "158574 0.43828500000000004 0.989003\n",
      "158603 0.43828500000000004 0.99433\n",
      "158694 0.43828500000000004 0.997734\n",
      "158749 0.43828500000000004 0.994421\n",
      "158751 0.43828500000000004 0.965948\n",
      "158777 0.43828500000000004 1.0\n",
      "158920 0.43828500000000004 0.992439\n",
      "158931 0.43828500000000004 0.975653\n",
      "158977 0.43828500000000004 0.960541\n",
      "159046 0.43828500000000004 0.965011\n",
      "159117 0.43828500000000004 0.994545\n",
      "159170 0.43828500000000004 0.989659\n",
      "159287 0.43828500000000004 0.989659\n",
      "159352 0.43828500000000004 0.973705\n",
      "159410 0.43828500000000004 0.990026\n",
      "159456 0.43828500000000004 0.991554\n",
      "159545 0.43828500000000004 0.986386\n",
      "159717 0.43828500000000004 0.993938\n",
      "159729 0.43828500000000004 0.960071\n",
      "159894 0.43828500000000004 0.982684\n",
      "159928 0.43828500000000004 0.993983\n",
      "160010 0.43828500000000004 0.997553\n",
      "160059 0.43828500000000004 0.987184\n",
      "160094 0.43828500000000004 0.965425\n",
      "160105 0.43828500000000004 0.970917\n",
      "160110 0.43828500000000004 0.96192\n",
      "160119 0.43828500000000004 0.988543\n",
      "160223 0.43828500000000004 0.986386\n",
      "160296 0.43828500000000004 0.964304\n",
      "160366 0.43828500000000004 0.999314\n",
      "160382 0.43828500000000004 0.985728\n",
      "160493 0.43828500000000004 0.992368\n",
      "160525 0.43828500000000004 0.991726\n",
      "160542 0.43828500000000004 0.994097\n",
      "160559 0.43828500000000004 0.975712\n",
      "160634 0.43828500000000004 0.963687\n",
      "160657 0.43828500000000004 0.985735\n",
      "160701 0.43828500000000004 0.973944\n",
      "160712 0.43828500000000004 0.9938\n",
      "160776 0.43828500000000004 0.995918\n",
      "160865 0.43828500000000004 0.993009\n",
      "160949 0.43828500000000004 0.970749\n",
      "161107 0.43828500000000004 0.968012\n",
      "161109 0.43828500000000004 0.993009\n",
      "161185 0.43828500000000004 0.961135\n",
      "161290 0.43828500000000004 0.991866\n",
      "161312 0.43828500000000004 0.99871\n",
      "161376 0.43828500000000004 0.9938\n",
      "161514 0.43828500000000004 0.99837\n",
      "161640 0.43828500000000004 0.997693\n",
      "161658 0.43828500000000004 0.991866\n",
      "162019 0.43828500000000004 0.997553\n",
      "162027 0.43828500000000004 0.981753\n",
      "162074 0.43828500000000004 0.983464\n",
      "162099 0.43828500000000004 0.989659\n",
      "162132 0.43828500000000004 0.982851\n",
      "162291 0.43828500000000004 0.972119\n",
      "162325 0.43828500000000004 0.965622\n",
      "162350 0.43828500000000004 0.986386\n",
      "162421 0.43828500000000004 0.987184\n",
      "162435 0.43828500000000004 0.954316\n",
      "162444 0.43828500000000004 0.965622\n",
      "162457 0.43828500000000004 0.975227\n",
      "162475 0.43828500000000004 0.972119\n",
      "162524 0.43828500000000004 0.999988\n",
      "162532 0.43828500000000004 0.965359\n",
      "162844 0.43828500000000004 0.97773\n",
      "162857 0.43828500000000004 0.972806\n",
      "162926 0.43828500000000004 0.993343\n",
      "162951 0.43828500000000004 0.993009\n",
      "163011 0.43828500000000004 0.985574\n",
      "163312 0.43828500000000004 0.995562\n",
      "163334 0.43828500000000004 0.970917\n",
      "163342 0.43828500000000004 0.983131\n",
      "163348 0.43828500000000004 0.994421\n",
      "163359 0.43828500000000004 0.989659\n",
      "163381 0.43828500000000004 0.973566\n",
      "163571 0.43828500000000004 0.992266\n",
      "163599 0.43828500000000004 0.971264\n",
      "163662 0.43828500000000004 0.983962\n",
      "163963 0.43828500000000004 0.957033\n",
      "164006 0.43828500000000004 0.976675\n",
      "164009 0.43828500000000004 0.991866\n",
      "164052 0.43828500000000004 0.998525\n",
      "164110 0.43828500000000004 0.991726\n",
      "164149 0.43828500000000004 0.992368\n",
      "164159 0.43828500000000004 0.957033\n",
      "164206 0.43828500000000004 0.987799\n",
      "164324 0.43828500000000004 0.96192\n",
      "164346 0.43828500000000004 0.99336\n",
      "164538 0.43828500000000004 0.984811\n",
      "165256 0.43828500000000004 0.973001\n",
      "165342 0.43828500000000004 0.992786\n",
      "165557 0.43828500000000004 0.990343\n",
      "165733 0.43828500000000004 0.970764\n",
      "165801 0.43828500000000004 0.969344\n",
      "165853 0.43828500000000004 0.962726\n",
      "165947 0.43828500000000004 0.991554\n",
      "165989 0.43828500000000004 0.991886\n",
      "166061 0.43828500000000004 0.973566\n",
      "166286 0.43828500000000004 1.0\n",
      "166287 0.43828500000000004 0.98473\n",
      "166314 0.43828500000000004 0.961135\n",
      "166339 0.43828500000000004 0.989948\n",
      "166399 0.43828500000000004 0.984811\n",
      "166452 0.43828500000000004 0.970141\n",
      "166678 0.43828500000000004 0.989003\n",
      "166786 0.43828500000000004 0.999917\n",
      "166798 0.43828500000000004 0.994883\n",
      "166820 0.43828500000000004 0.972119\n",
      "166881 0.43828500000000004 0.971264\n",
      "166939 0.43828500000000004 0.968073\n",
      "166995 0.43828500000000004 0.987516\n",
      "167026 0.43828500000000004 0.993009\n",
      "167071 0.43828500000000004 0.989424\n",
      "167130 0.43828500000000004 0.957033\n",
      "167203 0.43828500000000004 0.992786\n",
      "167210 0.43828500000000004 0.992439\n",
      "167227 0.43828500000000004 0.970749\n",
      "167340 0.43828500000000004 0.967767\n",
      "167447 0.43828500000000004 0.965622\n",
      "167457 0.43828500000000004 0.962726\n",
      "167672 0.43828500000000004 0.972119\n",
      "167685 0.43828500000000004 0.961557\n",
      "167698 0.43828500000000004 0.968467\n",
      "167796 0.43828500000000004 0.992198\n",
      "167842 0.43828500000000004 0.983623\n",
      "167990 0.43828500000000004 0.95519\n",
      "168038 0.43828500000000004 0.987184\n",
      "168091 0.43828500000000004 0.961557\n",
      "168186 0.43828500000000004 0.988122\n",
      "168223 0.43828500000000004 0.986386\n",
      "168228 0.43828500000000004 0.998101\n",
      "168245 0.43828500000000004 0.954531\n",
      "168368 0.43828500000000004 0.993114\n",
      "168467 0.43828500000000004 0.987184\n",
      "168593 0.43828500000000004 0.960071\n",
      "168602 0.43828500000000004 0.986681\n",
      "168624 0.43828500000000004 0.956727\n",
      "168748 0.43828500000000004 0.993508\n",
      "168838 0.43828500000000004 0.967843\n",
      "168877 0.43828500000000004 0.986386\n",
      "168906 0.43828500000000004 0.962637\n",
      "168948 0.43828500000000004 0.962867\n",
      "169058 0.43828500000000004 0.991349\n",
      "169081 0.43828500000000004 0.952897\n",
      "169097 0.43828500000000004 0.991349\n",
      "169190 0.43828500000000004 0.990672\n",
      "169288 0.43828500000000004 0.980178\n",
      "169290 0.43828500000000004 0.98969\n",
      "169309 0.43828500000000004 0.98199\n",
      "169334 0.43828500000000004 0.982293\n",
      "169408 0.43828500000000004 0.998444\n",
      "169524 0.43828500000000004 0.972119\n",
      "169634 0.43828500000000004 0.998678\n",
      "169661 0.43828500000000004 0.963687\n",
      "169813 0.43828500000000004 0.989067\n",
      "169820 0.43828500000000004 0.993009\n",
      "169841 0.43828500000000004 0.990736\n",
      "169880 0.43828500000000004 0.972596\n",
      "170011 0.43828500000000004 0.991349\n",
      "170023 0.43828500000000004 0.975712\n",
      "170038 0.43828500000000004 0.960071\n",
      "170091 0.43828500000000004 0.979693\n",
      "170136 0.43828500000000004 0.990343\n",
      "170172 0.43828500000000004 0.992786\n",
      "170254 0.43828500000000004 0.997429\n",
      "170327 0.43828500000000004 0.991579\n",
      "170452 0.43828500000000004 0.992229\n",
      "170487 0.43828500000000004 0.975712\n",
      "170565 0.43828500000000004 0.985735\n",
      "170601 0.43828500000000004 0.999147\n",
      "170698 0.43828500000000004 0.991349\n",
      "170718 0.43828500000000004 0.998444\n",
      "170935 0.43828500000000004 0.965359\n",
      "171045 0.43828500000000004 0.98588\n",
      "171074 0.43828500000000004 0.957033\n",
      "171107 0.43828500000000004 0.994545\n",
      "171302 0.43828500000000004 0.978156\n",
      "171354 0.43828500000000004 0.970749\n",
      "171457 0.43828500000000004 0.997786\n",
      "171486 0.43828500000000004 0.985454\n",
      "171503 0.43828500000000004 0.97162\n",
      "171534 0.43828500000000004 0.999021\n",
      "171610 0.43828500000000004 0.985735\n",
      "171890 0.43828500000000004 0.952897\n",
      "171941 0.43828500000000004 0.997693\n",
      "171958 0.43828500000000004 0.99873\n",
      "171983 0.43828500000000004 0.997098\n",
      "172172 0.43828500000000004 0.964304\n",
      "172247 0.43828500000000004 0.998257\n",
      "172278 0.43828500000000004 0.972806\n",
      "172387 0.43828500000000004 0.998909\n",
      "172547 0.43828500000000004 0.9998\n",
      "172572 0.43828500000000004 0.990157\n",
      "172614 0.43828500000000004 0.994888\n",
      "172643 0.43828500000000004 0.993527\n",
      "172835 0.43828500000000004 0.970917\n",
      "172883 0.43828500000000004 1.0\n",
      "173053 0.43828500000000004 0.990343\n",
      "173152 0.43828500000000004 0.954936\n",
      "173155 0.43828500000000004 0.990343\n",
      "173270 0.43828500000000004 0.966972\n",
      "173350 0.43828500000000004 0.953884\n",
      "173382 0.43828500000000004 0.96192\n",
      "173453 0.43828500000000004 0.999931\n",
      "173472 0.43828500000000004 0.952897\n",
      "173475 0.43828500000000004 0.996478\n",
      "173560 0.43828500000000004 0.953884\n",
      "173595 0.43828500000000004 0.993724\n",
      "173662 0.43828500000000004 0.993837\n",
      "173689 0.43828500000000004 0.992293\n",
      "173706 0.43828500000000004 0.954936\n",
      "173894 0.43828500000000004 0.991302\n",
      "173898 0.43828500000000004 0.993009\n",
      "173941 0.43828500000000004 1.0\n",
      "174035 0.43828500000000004 0.99433\n",
      "174068 0.43828500000000004 0.999486\n",
      "174134 0.43828500000000004 0.995762\n",
      "174207 0.43828500000000004 0.996307\n",
      "174249 0.43828500000000004 0.997999\n",
      "174254 0.43828500000000004 0.975227\n",
      "174333 0.43828500000000004 0.998909\n",
      "174568 0.43828500000000004 0.993527\n",
      "174678 0.43828500000000004 0.998525\n",
      "174750 0.43828500000000004 0.96192\n",
      "174755 0.43828500000000004 0.990736\n",
      "174839 0.43828500000000004 0.999345\n",
      "175067 0.43828500000000004 0.990343\n",
      "175091 0.43828500000000004 0.989659\n",
      "175142 0.43828500000000004 0.995938\n",
      "175226 0.43828500000000004 0.973001\n",
      "175240 0.43828500000000004 0.962637\n",
      "175266 0.43828500000000004 0.983131\n",
      "175294 0.43828500000000004 0.999993\n",
      "175356 0.43828500000000004 0.982302\n",
      "175421 0.43828500000000004 0.999719\n",
      "175455 0.43828500000000004 0.967767\n",
      "175465 0.43828500000000004 0.970749\n",
      "175508 0.43828500000000004 0.999697\n",
      "175638 0.43828500000000004 0.968073\n",
      "175795 0.43828500000000004 0.961557\n",
      "175817 0.43828500000000004 0.99837\n",
      "175821 0.43828500000000004 0.973566\n",
      "175864 0.43828500000000004 0.996931\n",
      "175871 0.43828500000000004 0.966972\n",
      "175901 0.43828500000000004 0.989221\n",
      "176107 0.43828500000000004 1.0\n",
      "176118 0.43828500000000004 0.982302\n",
      "176148 0.43828500000000004 0.991902\n",
      "176172 0.43828500000000004 1.0\n",
      "176195 0.43828500000000004 0.988543\n",
      "176198 0.43828500000000004 0.994888\n",
      "176204 0.43828500000000004 0.960071\n",
      "176268 0.43828500000000004 0.987516\n",
      "176345 0.43828500000000004 0.993938\n",
      "176377 0.43828500000000004 0.998257\n",
      "176393 0.43828500000000004 0.972119\n",
      "176401 0.43828500000000004 0.970749\n",
      "176445 0.43828500000000004 0.95483\n",
      "176617 0.43828500000000004 0.998808\n",
      "176668 0.43828500000000004 0.997553\n",
      "176801 0.43828500000000004 0.972119\n",
      "176811 0.43828500000000004 0.964117\n",
      "177010 0.43828500000000004 0.987184\n",
      "177130 0.43828500000000004 0.994883\n",
      "177302 0.43828500000000004 0.988543\n",
      "177322 0.43828500000000004 0.962637\n",
      "177355 0.43828500000000004 0.999021\n",
      "177374 0.43828500000000004 0.999002\n",
      "177384 0.43828500000000004 1.0\n",
      "177439 0.43828500000000004 0.972119\n",
      "177655 0.43828500000000004 0.998444\n",
      "177848 0.43828500000000004 0.960071\n",
      "177889 0.43828500000000004 0.972119\n",
      "177963 0.43828500000000004 0.982684\n",
      "177967 0.43828500000000004 0.961135\n",
      "178008 0.43828500000000004 0.997553\n",
      "178042 0.43828500000000004 0.972119\n",
      "178187 0.43828500000000004 0.960541\n",
      "178203 0.43828500000000004 0.981023\n",
      "178242 0.43828500000000004 0.957937\n",
      "178268 0.43828500000000004 0.985728\n",
      "178354 0.43828500000000004 0.996478\n",
      "178393 0.43828500000000004 0.952897\n",
      "178636 0.43828500000000004 0.982302\n",
      "178668 0.43828500000000004 0.997429\n",
      "178712 0.43828500000000004 0.987799\n",
      "178785 0.43828500000000004 0.997704\n",
      "178838 0.43828500000000004 0.999095\n",
      "178844 0.43828500000000004 0.991118\n",
      "178855 0.43828500000000004 0.970369\n",
      "178908 0.43828500000000004 0.965786\n",
      "179078 0.43828500000000004 0.960354\n",
      "179310 0.43828500000000004 0.972119\n",
      "179331 0.43828500000000004 0.980539\n",
      "179343 0.43828500000000004 0.974758\n",
      "179772 0.43828500000000004 0.960541\n",
      "179795 0.43828500000000004 0.993983\n",
      "179807 0.43828500000000004 0.993114\n",
      "179822 0.43828500000000004 0.988543\n",
      "179848 0.43828500000000004 0.998854\n",
      "179899 0.43828500000000004 0.965786\n",
      "180018 0.43828500000000004 0.997553\n",
      "180142 0.43828500000000004 0.970917\n",
      "180336 0.43828500000000004 0.965359\n",
      "180463 0.43828500000000004 0.999021\n",
      "180530 0.43828500000000004 0.991604\n",
      "180556 0.43828500000000004 0.975108\n",
      "180645 0.43828500000000004 0.985838\n",
      "180900 0.43828500000000004 0.999353\n",
      "180938 0.43828500000000004 0.988543\n",
      "181169 0.43828500000000004 0.95519\n",
      "181278 0.43828500000000004 0.973566\n",
      "181337 0.43828500000000004 0.973566\n",
      "181456 0.43828500000000004 0.989659\n",
      "181489 0.43828500000000004 0.994883\n",
      "181495 0.43828500000000004 0.995762\n",
      "181504 0.43828500000000004 0.992289\n",
      "181595 0.43828500000000004 0.997553\n",
      "181596 0.43828500000000004 0.994994\n",
      "181638 0.43828500000000004 0.986681\n",
      "181734 0.43828500000000004 0.980539\n",
      "181916 0.43828500000000004 0.999987\n",
      "182216 0.43828500000000004 0.979276\n",
      "182271 0.43828500000000004 0.972728\n",
      "182432 0.43828500000000004 0.981023\n",
      "182440 0.43828500000000004 0.972596\n",
      "182679 0.43828500000000004 0.999095\n",
      "182700 0.43828500000000004 0.962637\n",
      "182709 0.43828500000000004 0.999486\n",
      "183016 0.43828500000000004 0.959748\n",
      "183203 0.43828500000000004 0.965011\n",
      "183367 0.43828500000000004 0.952897\n",
      "183369 0.43828500000000004 0.997479\n",
      "183422 0.43828500000000004 1.0\n",
      "183702 0.43828500000000004 0.973566\n",
      "183733 0.43828500000000004 0.998257\n",
      "183768 0.43828500000000004 0.974514\n",
      "183884 0.43828500000000004 0.972119\n",
      "183920 0.43828500000000004 0.998283\n",
      "183932 0.43828500000000004 0.989659\n",
      "184064 0.43828500000000004 0.981646\n",
      "184093 0.43828500000000004 0.985735\n",
      "184097 0.43828500000000004 0.998671\n",
      "184112 0.43828500000000004 0.970917\n",
      "184151 0.43828500000000004 0.997553\n",
      "184242 0.43828500000000004 0.991706\n",
      "184264 0.43828500000000004 0.96192\n",
      "184332 0.43828500000000004 0.984826\n",
      "184354 0.43828500000000004 0.980539\n",
      "184364 0.43828500000000004 0.993837\n",
      "184480 0.43828500000000004 0.999115\n",
      "184488 0.43828500000000004 0.965425\n",
      "184642 0.43828500000000004 0.972119\n",
      "184667 0.43828500000000004 0.998678\n",
      "184671 0.43828500000000004 0.982302\n",
      "184736 0.43828500000000004 0.952897\n",
      "184752 0.43828500000000004 0.965425\n",
      "184767 0.43828500000000004 0.961135\n",
      "184820 0.43828500000000004 0.960541\n",
      "184832 0.43828500000000004 0.978324\n",
      "184871 0.43828500000000004 0.976002\n",
      "184894 0.43828500000000004 0.986386\n",
      "185000 0.43828500000000004 0.993009\n",
      "185036 0.43828500000000004 0.972714\n",
      "185075 0.43828500000000004 0.957033\n",
      "185097 0.43828500000000004 0.953448\n",
      "185170 0.43828500000000004 0.964304\n",
      "185185 0.43828500000000004 0.99433\n",
      "185223 0.43828500000000004 0.982302\n",
      "185235 0.43828500000000004 0.992676\n",
      "185318 0.43828500000000004 0.966972\n",
      "185364 0.43828500000000004 0.965948\n",
      "185448 0.43828500000000004 0.992293\n",
      "185460 0.43828500000000004 0.967843\n",
      "185490 0.43828500000000004 0.998444\n",
      "185524 0.43828500000000004 0.998444\n",
      "185530 0.43828500000000004 0.993508\n",
      "185559 0.43828500000000004 0.991349\n",
      "185787 0.43828500000000004 0.957937\n",
      "185843 0.43828500000000004 0.998257\n",
      "185958 0.43828500000000004 0.988543\n",
      "186277 0.43828500000000004 0.965011\n",
      "186426 0.43828500000000004 0.966639\n",
      "186431 0.43828500000000004 0.994545\n",
      "186444 0.43828500000000004 0.997693\n",
      "186667 0.43828500000000004 0.991349\n",
      "186675 0.43828500000000004 0.967767\n",
      "186751 0.43828500000000004 0.952897\n",
      "186860 0.43828500000000004 0.960071\n",
      "186944 0.43828500000000004 0.95519\n",
      "187346 0.43828500000000004 0.994883\n",
      "187424 0.43828500000000004 0.952897\n",
      "187460 0.43828500000000004 0.994421\n",
      "187486 0.43828500000000004 0.982684\n",
      "187557 0.43828500000000004 0.995762\n",
      "187656 0.43828500000000004 0.977482\n",
      "187682 0.43828500000000004 0.990157\n",
      "187724 0.43828500000000004 0.992676\n",
      "187762 0.43828500000000004 0.991866\n",
      "187839 0.43828500000000004 0.970369\n",
      "187853 0.43828500000000004 0.985735\n",
      "188026 0.43828500000000004 0.997704\n",
      "188084 0.43828500000000004 0.95559\n",
      "188088 0.43828500000000004 0.95519\n",
      "188119 0.43828500000000004 0.988145\n",
      "188134 0.43828500000000004 0.997553\n",
      "188197 0.43828500000000004 0.989659\n",
      "188250 0.43828500000000004 0.954423\n"
     ]
    }
   ],
   "source": [
    "# Get median of cont7 values\n",
    "\n",
    "cont7_median = train.cont7.median()\n",
    "print(cont7_median)\n",
    "\n",
    "# Create series with values from train.cont7 but replacing outliers with median\n",
    "new_cont7 = (cont7_median - train.cont7) * outliers.cont7 + train.cont7\n",
    "\n",
    "# Check that count of new values equal to median value is as expected (minus existing median values)\n",
    "print(new_cont7[new_cont7==cont7_median].shape[0] - train.cont7[train.cont7==cont7_median].shape[0])\n",
    "\n",
    "# Confirmed: count matches # of outliers found\n",
    "\n",
    "# display new and old values if new value != old value\n",
    "for ndx, valu in new_cont7.iteritems():\n",
    "    if valu != train.cont7.iloc[ndx]:\n",
    "        print(ndx, valu, train.cont7.iloc[ndx])\n",
    "\n",
    "# Confirmed: old values look like outliers and new value is the median value\n",
    "\n",
    "# Replace cont7 in train data set with new values\n",
    "train.cont7 = new_cont7\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {},
   "outputs": [],
   "source": [
    "# split data into train and test dfs\n",
    "X_train, X_test, y_train, y_test = train_test_split(train, train_loss, test_size=.2, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0:01:14.141183\n",
      "The mae of the model prediction is:  1140.23833551\n",
      "\n",
      "Best params:\n",
      "{'learning_rate': 0.1, 'min_data_in_leaf': 125, 'n_estimators': 500, 'num_leaves': 16}\n"
     ]
    }
   ],
   "source": [
    "start=datetime.now()\n",
    "lgbm_reg_best.fit(X_train, y_train)\n",
    "stop=datetime.now()\n",
    "print(stop-start)\n",
    "\n",
    "y_pred=lgbm_reg_best.predict(X_test)\n",
    "\n",
    "# adjust the predictions (due to adjustment of loss)\n",
    "adj_y_test = np.exp(y_test) - shift\n",
    "adj_y_pred = np.exp(y_pred) - shift\n",
    "\n",
    "#print('The mae of the model prediction is: ', mean_absolute_error(y_test, y_pred))\n",
    "print('The mae of the model prediction is: ', mean_absolute_error(adj_y_test, adj_y_pred))\n",
    "\n",
    "print('\\nBest params:\\n{}'.format(lgbm_reg_best.best_params_))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {},
   "outputs": [],
   "source": [
    "# no improvement - this transformation not included in final model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "metadata": {},
   "outputs": [],
   "source": [
    "# try removing outlier rows\n",
    "\n",
    "out_cont7_ndx = outliers[outliers.cont7 == True].index\n",
    "\n",
    "train = train.drop(out_cont7_ndx)\n",
    "train_loss = train_loss.drop(out_cont7_ndx)\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "185659 185659\n"
     ]
    }
   ],
   "source": [
    "print(train.shape[0], train_loss.shape[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "metadata": {},
   "outputs": [],
   "source": [
    "# split data into train and test dfs\n",
    "X_train, X_test, y_train, y_test = train_test_split(train, train_loss, test_size=.2, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0:01:13.514357\n",
      "The mae of the model prediction is:  1128.01689189\n",
      "\n",
      "Best params:\n",
      "{'learning_rate': 0.1, 'min_data_in_leaf': 125, 'n_estimators': 500, 'num_leaves': 16}\n"
     ]
    }
   ],
   "source": [
    "start=datetime.now()\n",
    "lgbm_reg_best.fit(X_train, y_train)\n",
    "stop=datetime.now()\n",
    "print(stop-start)\n",
    "\n",
    "y_pred=lgbm_reg_best.predict(X_test)\n",
    "\n",
    "# adjust the predictions (due to adjustment of loss)\n",
    "adj_y_test = np.exp(y_test) - shift\n",
    "adj_y_pred = np.exp(y_pred) - shift\n",
    "\n",
    "#print('The mae of the model prediction is: ', mean_absolute_error(y_test, y_pred))\n",
    "print('The mae of the model prediction is: ', mean_absolute_error(adj_y_test, adj_y_pred))\n",
    "\n",
    "print('\\nBest params:\\n{}'.format(lgbm_reg_best.best_params_))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "cont7    0.743485\n",
       "dtype: float64"
      ]
     },
     "execution_count": 115,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# examine skew of remaining continuous features\n",
    "# Code adapated from here: https://www.kaggle.com/snmateen/simple-eda-feature-transformations/notebook\n",
    "\n",
    "sku = train[cols_cont_feat].skew()\n",
    "sku[abs(sku) > .7]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY0AAAEWCAYAAACaBstRAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nJGn/B3EP+ix1sCH4AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2730604b828>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot deviation from normal distribution\n",
    "# Code adapated from here: https://www.kaggle.com/snmateen/simple-eda-feature-transformations/notebook\n",
    "\n",
    "ax=plt.gca()\n",
    "stats.probplot(train.cont7, dist=stats.norm, plot=ax, rvalue=True)\n",
    "ax.set_title('Probplot of feature')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAEKCAYAAADpfBXhAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nfIQAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2730604b2b0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Code adapated from here: https://www.kaggle.com/snmateen/simple-eda-feature-transformations/notebook\n",
    "\n",
    "sns.distplot(train.cont7, bins=50)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "metadata": {},
   "outputs": [],
   "source": [
    "# use boxcox to transform highly-skewed feature, cont7 - that is, normalize it\n",
    "# Based on example code from here: https://www.kaggle.com/snmateen/simple-eda-feature-transformations/notebook\n",
    "\n",
    "xcf, _ = stats.boxcox(train.cont7)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAEWCAYAAABxMXBSAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4wLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvpW3flQAAIABJREFUeJzt3Xd4VGXax/Hvj7aKBRsoRUBXhAAC\nahTsulZQsaIitrXwCiI2VnCxK8paUJSisK6i4oodFFRABFZAIUiHoChIVVFRkChS7vePc4KTMJPM\nJJlMMrk/15Vr5pw55Z6TZO55ynkemRnOOedcvCqlOgDnnHPliycO55xzCfHE4ZxzLiGeOJxzziXE\nE4dzzrmEeOJwzjmXEE8crlgk3Svp5SLue6KklSUdU4xz7StpsqQNkh4vjXMmQtKDkn6Q9G2qY8mv\nsN+xpC6SvpP0q6S9SzM2lxqeOCogScsk/Rb+o38n6XlJu6Y6roJIekHSg8U4RGfgB2B3M7stCccv\nMkn7A7cBTc1sv1TEUFSSqgL9gNPMbFcz+7EYx2ooySRVKbkIXTJ44qi4zjazXYHDgCOAO/NvoEC6\n/I00ABZaEe94TfKHWQPgRzP7PtEdy8CH7L7ATsCCFMeRbn+vZZpf5ArOzFYB7wPNASRNlNRH0hQg\nBzhQUh1JoyT9JGmJpOvyHWYnSSPCaqDPJbXMfSEs3dwhaaGkdWHpZqdosUjKCM//s6QFktqH6zsD\nnYDbw1LSuzH2P1rSDEm/hI9Hh+tfAK6M2P+UfPtFPX4Ye09Jc4GNkqpI6iXpq/C9LpR0XsRxrpL0\niaTHwve6VFLbfK9/He67VFKnMJZxQJ3w3C+E27YPr8HP4TXJyHdN88e1TNI/JM2VtFHSc2H13Pvh\n+cZL2jPiGG0kTQ2PP0fSiRGvHSBpUrjfOGCfGNf7YGBxuPizpAnh+iaSxoV/L4slXRSxz5mSZkla\nL2mFpHsjDjk54li/SjpK+arJ8pdKYvy91gjf/xpJqxRUA1aO9h5cEZmZ/1SwH2AZcEr4fH+Cb4sP\nhMsTgeVAM6AKUBWYBAwi+GbZClgLnBxufy+wGbgw3LYHsBSoGnGu+eF59gKmAA+Gr50IrAyfVwWW\nAP8EqgF/AzYAjcPXX8jdL8Z72gtYB1wext0xXN47zv13eD2MfXYY+87hug5AHYIvXRcDG4Ha4WtX\nhdfiOqAy0AVYDQjYBVgf8X5qA83yX4dw+eDwuKeG1+X28NpUKyCuZcCnBCWAusD3wOfAocBfgAnA\nPeG2dYEfgXbh+zg1XK4Zvj6NoPrpL8Dx4e/h5RjXrSFgQJVweRdgBfD38PdwGEEVYeR7PSQ8bwvg\nO+DcaMeK+Pt6uYDzTWTHv9d3gGfDWGoB04H/S/X/XTr9eImj4npH0s/AJwSJ4aGI114wswVmtgXY\nDzgW6Glmv5vZbODfBB/QuWaa2RtmtpngA2cnoE3E6wPMbIWZ/QT0IfhQz68NsCvQ18z+MLMJwHsx\nto3mTOBLM3vJzLaY2X+BbODsOPeP5akw9t8AzOx1M1ttZtvMbATwJXBkxPbfmNlQM9sKDCNIEPuG\nr20Dmkva2czWmFms6p2LgdFmNi68po8BOwNHx4or9LSZfWdBKfJ/wGdmNsvMNgFvEyQRgMuAMWY2\nJnwf44AsoJ2k+gRVl3eZ2SYzmwxELeHFcBawzMyeD38PnwNvEnyxwMwmmtm88Lxzgf8CJyRw/Ggi\n/173AtoCN5vZRguq/54ALinmOVwETxwV17lmtoeZNTCzrvk+gFZEPK8D/GRmGyLWfUPwrXWH7c1s\nG7Ay3C/a8b7J91rkeVaE+8c6T0HqhNtHSmT/WCJjR9IVkmaHVTw/E1TxRVblbO8VZWY54dNdzWwj\nQUK4HlgjabSkJjHOmee9hNdkBTGueYTvIp7/FmU5twNEA6BD7nsI38exBEmuDrAujDdX/utakAZA\n63zH7kTwBQRJrSV9LGmtpF8IrkfUqrAERF6LBgSljjUR53+WoOThSognDhdNZAPyamAvSbtFrKsP\nrIpY3j/3iYLGyXrhfju8Hu4b+VrkefZX3sbNyPMU1qi9muBDI1L+OAsS6/jb10tqAAwFuhFUge1B\nUA2nuE5g9qGZnUrwAZ0dHiuaPO9FkgiuYeR7Kc6w1iuAl8IvDrk/u5hZX2ANsKekXSK2r5/gsSfl\nO/auZtYlfP0VYBSwv5nVAJ7hz+sX7T1tBKpHLEfrdRa53wpgE7BPxPl3N7NmCbwHVwhPHK5AZrYC\nmAo8LGknSS2Aa4DhEZsdLun8sMHyZoJ/3E8jXr9BUj1JexG0YYyIcqrPCD4kbpdUNWysPRt4NXz9\nO+DAAkIdAxws6dKwsfhioClBdVc8Cjs+BHXmRtDGg6S/E3YqKEzYUN0+/EDeBPwKbI2x+WvAmZJO\nVtDd9bZwn6nxnCsOLwNnSzpdUuXw93qipHpm9g1BtdV9kqpJOpbEqvveI/g9XB7+HqtKOiKicX83\nghLs75KOBC6N2HctQXVe5O9hNnC8pPqSagB3FHRyM1sDjAUel7S7pEqS/iqpuNVhLoInDhePjgSN\nkqsJ6srvCevFc40kqIbJbZw+P6ybz/UKwT/z1+HPDvdLmNkfQHuC+ukfCBrjrzCz7HCT54CmYfXD\nO1H2/5Ggfv02gobe24GzzOyHON9jgccPz7EQeJyg8fg7gkbeKXEev1IY22rgJ4J6/a4xzrOYoB3i\naYJrcTZB9+k/4jxXgcIvA+cQJPG1BN/S/8GfnweXAq3DOO8BXkzg2BuA0wjaFFYTVN39i6ChHYL3\nfL+kDcDdBEkyd98cgjawKeHvoU34dzYCmAvMJL4vAlcQdLBYSPA3+QZBKc+VEJn5RE4ueSQtA641\ns/GpjsU5VzK8xOGccy4hnjicc84lxKuqnHPOJcRLHM455xKS6gHSkmKfffaxhg0bpjoM55wrN2bO\nnPmDmdWMZ9u0TBwNGzYkKysr1WE451y5ISnuEQK8qso551xCPHE455xLiCcO55xzCfHE4ZxzLiGe\nOJxzziXEE4dzzrmEeOJwzjmXEE8czjmXDiZPhkceKZVTeeJwzrny7Pvv4cor4YQT4NlnYePGwvcp\nJk8czjlXHm3bFiSKJk3glVfgjjtg3jzYZZfC9y2mtBxyxDnn0tqsWdClC3z2WVDSGDQImjYttdN7\nicM558qL9evh5pshMxO+/hpefBE+/rhUkwZ4icM558o+M3j9dbjlFlizBv7v/+Chh2DPPVMSjpc4\nnHOuLFuyBM44Ay6+GPbdF6ZNg8GDU5Y0wBOHc86VTb//DvfdB82bB8mif3+YPh1at051ZF5V5Zxz\nZc64cdC1a1DauPhi6NcP6tRJdVTbeYnDOefKitWrg0Rx2mnB8tix8OqrZSppQIoTh6QzJC2WtERS\nryivXyVpraTZ4c+1qYjTOeeSasuWoCqqSRMYORLuvTe4J+PUU1MdWVQpq6qSVBkYCJwKrARmSBpl\nZgvzbTrCzLqVeoDOOVcaPvssuCdj1iw4/XQYMAAOOijVURUolSWOI4ElZva1mf0BvAqck8J4nHOu\n9KxbB9dfD0cdBd99B6+9Bu+/X+aTBqQ2cdQFVkQsrwzX5XeBpLmS3pC0f+mE5pxzSWIW3LjXuDEM\nHQo33QSLFkGHDiClOrq4pDJxRLtClm/5XaChmbUAxgPDYh5M6iwpS1LW2rVrSzBM55wrIQsXwkkn\nBYMSHnggzJwJTzwBu++e6sgSksrEsRKILEHUA1ZHbmBmP5rZpnBxKHB4rIOZ2RAzyzSzzJo1a5Z4\nsM45V2Q5OcEghC1bwty5weCEU6dCq1apjqxIUpk4ZgCNJB0gqRpwCTAqcgNJtSMW2wOLSjE+55wr\nvnffDcaS6tsXOnWC7Gzo3Bkqld+7IVLWq8rMtkjqBnwIVAb+Y2YLJN0PZJnZKKC7pPbAFuAn4KpU\nxeuccwlZvhy6dw+61zZtCpMmwfHHpzqqEiGz/M0K5V9mZqZlZWWlOgznXEW0eXPQbnHffcHy3XcH\ngxNWq5bauAohaaaZZcazrQ854pxzJWXy5OCejIULoX17eOopaNAg1VGVuPJbyeacc2XF2rVw1VXB\npEq//hpUT40cmZZJAzxxOOdc0W3bBkOGBPdkDB8OvXr9WdpIY15V5ZxzRTF7dlAt9emnQaP34MGl\nPhNfqniJwznnErFhQ9DYffjh8NVXMGwYTJxYYZIGeInDOefiYwZvvBHM+b1mTXAvxkMPwV57pTqy\nUuclDuecK8ySJdC2LVx0EdSqFdz1/cwzFTJpgCcO55yLbdMmuP/+YPrWqVPhySdhxgxo0ybVkaWU\nV1U551w048cH07d++WVQ0ujXD+pGG8C74vESh3PORVqzBjp2DGbfM4MPP4QRIzxpRPDE4ZxzAFu3\nwtNPB9O3vvUW3HNPMH1r7vzfbjuvqnLOuenTg9n4Zs0KShoDB0KjRqmOqszyEodzruJaty64ia9N\nG/j226BK6sMPPWkUwhOHc67iMYOXXgqqpYYMCYY/z84OGsHLyfStqeRVVc65imXRoqC31MSJ0Lo1\nfPABHHpoqqMqV7zE4ZyrGHJy4J//DKZvnT07uIFv6lRPGkXgJQ7nXPp77z248UZYtgyuuAIefTS4\nA9wViZc4nHPpa/lyOO88OPts2HnnoHpq2DBPGsXkicM5l342bw5KFRkZQS+phx8OqqdOOCHVkaUF\nr6pyzqWXTz4JutjOnx+UNJ56Cho2THVUacVLHM659LB2LVx9NRx3HKxfD++8A6NGedJIAk8czrny\nbds2GDo0uCfjpZegZ89g+tZzzkl1ZGnLq6qcc+XXnDlBtdS0aUFJY/BgaNYs1VGlvZSWOCSdIWmx\npCWSekV5/S+SRoSvfyapYelH6ZwrczZsgFtvDaZv/fJLeOEFmDTJk0YpSVnikFQZGAi0BZoCHSXl\nn7T3GmCdmR0EPAH8q3SjdM6VKbnTt2ZkwBNPwDXXwOLFcOWVPlRIKUplieNIYImZfW1mfwCvAvkr\nJc8BhoXP3wBOlvyvw7kK6auvoF076NAB9tknqJ569tkKO31rKqUycdQFVkQsrwzXRd3GzLYAvwB7\nRzuYpM6SsiRlrV27NgnhOudSYtMmeOCBYPrWTz4JShpZWRV++tZUSmXiiFZysCJsE6w0G2JmmWaW\nWbNmzWIH55wrAz76CFq0gLvvDu7JyM6Gm2+GKrH79Ujl+6dy5eCxSpW8jw0bBmMzNmwIlSoFj8OH\nB+95+PDo65MllYljJbB/xHI9YHWsbSRVAWoAP5VKdM651FmzBi69FE45BbZu5QzeR6+/hurVLfSD\nt7zbti143Lo17+M33wSdxr75Jmjq+eYb6Nw5SCadO++4PpnJI5WJYwbQSNIBkqoBlwCj8m0zCrgy\nfH4hMMHMopY4nHNpYOtWeu8xgF/qNGHTf9/kPu5m56/m8SFnpDqyMiknJ5hOJCdnx/W9eyfvvCm7\nj8PMtkjqBnwIVAb+Y2YLJN0PZJnZKOA54CVJSwhKGpekKl7nXHLklhIymcFgutCHmYzjFG5gIF9y\ncGqDKwdySyT5LV+evHOm9AZAMxsDjMm37u6I578DHUo7Ludc8uSvTqrBz/ShN10YzLfsx8W8ymtc\nRPQmTpdf5crRk0f9+sk7pw854pxLuuhtEEYnXmYxjbmeZ3iaG8lgEa9xMZ404lO9etCeUb36juv7\n9EneeT1xOOeSJlaDdWOy+YiTeZnLWUZDjmAGN9Of9dQo/SDLmErhp3LlynkfGzQIRldp0CC4pg0a\nBO0bgwYFj/nXd+qUvBh9rCrnXImL1btpZ3LoTR/+waNsZBeuZzBDuY5tVE5qPOnepaZTp+Qmivw8\ncTjnSkS1asH8SbG0YzQD6MYBLONFLucfPMr37Fvk86V7MijLvKrKOVdsUuykUY8VvMn5jOYsfmNn\nTuRjruTFhJKG2Y4/LnU8cTjniqSwm+6qsJnbeIxFZHAGH3AHD9GK2UzixEKP7UmibPOqKudcwgq7\nQ/topvAM13MI83mXs+jOUyzjgJjbe3IoX7zE4ZyLW2HDeuzND/yba5jCsdTgF87lbdozKmrS8BJF\n+eWJwzlXqMIShtjG1TzHYhpzBS/yL24ng0WM5Fzy35PhyaL886oq51xM8QwaeAhzGUwXjmEqkzmO\nrgxiAc132M6TRfrwEodzbgfxjDS7Kxt4jNv4nMM4mC+4iuc5gUl5koZXR6UnTxzOuTwKL2UY5/Mm\ni8jgNvrxH66mCdkM4ypyq6WaNvVkkc4SqqqStCewv5nNTVI8zrkUiada6gC+5mlu5EzGMJuWdOB1\nPuWo7a97sqgYCi1xSJooaXdJewFzgOcl9Ut+aM650lJY0qjGJnrzIAtoxvFM5hb6kUmWJ40KKp6q\nqhpmth44H3jezA4HTkluWM650hBPW8ZJTGAOLXmQu3iPs8hgEU9yC1vDCgtvw6h44kkcVSTVBi4C\n3ktyPM65UhBPwtiXb3mZTkzgZKqymTN4n4t4nVXUAzxhVGTxJI77CWbp+8rMZkg6EPgyuWE555Ih\nnoRRia10ZSDZNOFC3uB+7qI587dP3+oJwxXaOG5mrwOvRyx/DVyQzKCccyUvnsbvw8liMF04gizG\nczJdGbR9+lZPFi5XPI3jB0v6SNL8cLmFpDuTH5pzrqQUljRq8DNP043pHEk9VnIJ/+VUxnnScFHF\nU1U1FLgD2AwQdsW9JJlBOedKTsFJw7iU4WTThC4MZgDdaEI2I7gEkFdLuajiuY+juplNV96/vi1J\nisc5V0IKK2U0JpuB3MDJTGA6R3Amo/mcw7e/7gnDxRJPieMHSX8FDEDShcCapEblnCuWgpLGTvzG\nA9zJXFpwODPpwiCOYtr2pOGlDFeYeEocNwBDgCaSVgFLgcuSGpVzrsgKShptGcMAunEgS3mJy+jB\nY9tn4vNk4eIVT6+qr4FTJO0CVDKzDcU9aXgX+gigIbAMuMjM1kXZbiswL1xcbmbti3tu59JVQQmj\nHit4kpu5gLdYRBNOYgITOQnwhOESV2jikHR3vmUAzOz+Ypy3F/CRmfWV1Ctc7hllu9/MrFUxzuNc\n2isoYVRhM915ivu4h8ps5Z/04TF6sJlqgCcNVzTxtHFsjPjZCrQlKCkUxznAsPD5MODcYh7PuQqp\noKRxNFOYyeE8Tg8mciJNWcjD/NOThiu2eKqqHo9clvQYMKqY593XzNaEx18jqVaM7XaSlEXQi6uv\nmb0T64CSOgOdAerXr1/M8Jwr2wpKGHvxI/+iJ9fyHCuox3m8xTv5ZuLzpOGKoygzAFYHDixsI0nj\ngf2ivNQ7gXPVN7PV4TAnEyTNM7Ovom1oZkMIGvHJzMz0fwuXtmIlDbGNq3iBR7idGvzCI/yD+7mb\njey6fRtPGK4kxNPGMY+wKy5QGahJMH5Vgcws5gi6kr6TVDssbdQGvo9xjNXh49eSJgKHAlETh3MV\nQayk0Zx5DKYLxzKF/3EsXRi8w/StnjRcSYmnxHFWxPMtwHdmVtwbAEcBVwJ9w8eR+TcIJ43KMbNN\nkvYBjgEeKeZ5nSu3oiWNXfiVe7iPW3iCn9mDv/MfhnEllq/50pOGK0kxE0fYZRYgf/fb3SVhZj8V\n47x9gdckXQMsBzqE58wErjeza4EM4FlJ2wga8fua2cJinNO5cil6KcM4j7fpz03sz0qGci296MtP\n7J13K08YLgkKKnHMJKiiivZna8TRzhGLmf0InBxlfRZwbfh8KnBIUc/hXDqIljQaspQBdONMxjCH\nFlzMCKZxdJ5tPGG4ZIqZOMzsgNIMxDmXV/6kUY1N9OAx7uRBtlCFW3mcp+i+fSa+XJ40XLLF1asq\nbG9oBOyUu87MJicrKOcquvxJ40Q+ZjBdaMJi3uACbubJ7TPx5fKE4UpLPL2qrgVuAuoBs4E2wDTg\nb8kNzbmKKTJp1OI7Huc2LmM4X3EgbRnDB7TdYR9PGq40xXPn+E3AEcA3ZnYSQZfYtUmNyrkKKHJa\n10pspQuDWExjLuI1HuBOmjPfk4YrE+KpqvrdzH6XhKS/mFm2pMZJj8y5CiSylHEYM3mG6zmCLD7i\nb3RlEF8Q/V/Ok4ZLhXhKHCsl7QG8A4yTNBJYndywnKs4cpPG7vzCU9y4ffrWjrzCKYyPmjR8zgyX\nSvGMVXVe+PReSR8DNYAPkhqVcxXAn6UM4xJepR+3si/fMZAbuJMHWU+NqPt5wnCpVtANgKOBV4B3\nzGwjgJlNKq3AnEtnuUnjYBYzkBs4hY+YQSZn8V6e6Vvz86ThyoKCqqqGEAw3skzSCEnnSqpWSnE5\nl7akYPrW+7mLubQgkyy6MpA2fOpJw5ULMROHmY00s45AfeAtgjGllkv6j6RTSytA59KJBGfwPvNp\nzl08yOt0oAnZDKYr26gccz9PGq4sKbRx3Mx+M7MRYVvHaQTdcb2Nw7kESFBPK3mdC3mfdmymKn/j\nIy7nZb6LOvtAwBvBXVlUaOKQtK+kGyVNIehZNRYKKE875/Kooi3cQj8WkcGZjKY3D9KSOXxcyD20\nnjBcWVVQ4/h1QEegMUFV1e1mNqW0AnOuvJPgKKYyky60ZC6jaceNPM3SOMYH9aThyrKCuuMeTTD8\n+Xgz21ZK8TiXFvbWjwyhF9fxb1ZQj/N5k7c5j+iDTf/JE4YrDwoaHffvpRmIc+mgkrZxJcPI5nb2\nZB2PcRv3cm+e6Vtj8aThyouizDnunMtHgmbMZxJdOI5PmMLRdGEw82gR1/6eNFx5Es+QI865Auyq\nX/kXtzObVmSwiKt5juP4nycNl7bimTo2qmJOHetcuScZ5/IOC7mJ+qzg31xDL/ryI/vEtb8nDFde\nxTt1bH1gXfh8D4J5wn2GQFdhHaClvMuNnMVo5nIIHfkvUzkm7v09abjyrNCpYyU9A4wyszHhclvg\nlNIJz7myQ4Kq/EEPHmMBD7KNStzGYzxFd7ZQNa5jeMJw6SCeNo4jcpMGgJm9D5yQvJCcK1tyJ1g6\nkY+ZQ0seojfv05YMFtGP2zxpuAonnsTxg6Q7JTWU1EBSb+DHZAfmXFkgBdO3vsjlfMzf+AubaMdo\nLuRNVrJ/XMfwYUNcuokncXQEagJvhz81w3XOpaXcEkZlbeV6BpNNEy5mBA/Sm+bM533axXUcTxgu\nXcUzyOFPZnYTcJyZHWZmNxe3R5WkDpIWSNomKbOA7c6QtFjSEkm9inNO5+KRO0/GoXzONI5iMF2Z\nxaG0YC538SC/UT2u43jCcOksnkEOj5a0EFgYLreUNKiY550PnA9MLuC8lYGBQFugKdBRUtNinte5\nHeSWMKRg+tb+dGcGR1Cf5XTiZU7mIxbTJK5jeSnDVQTxVFU9AZxO2K5hZnOA44tzUjNbZGaLC9ns\nSGCJmX1tZn8ArwLnFOe8zuUXOX3rxbxKNk3oxgAG04UmZPMKnShsfKntR/CE4SqIuO4cN7MV+VZt\nTUIs+dUFIs+7MlwXlaTOkrIkZa1duzbpwbnyK7KEAdCILxjLabxKR1ZRl9Z8xo0M4Bf2KPRYuSUM\nTxquIokncayQdDRgkqpJ6gEsKmwnSeMlzY/yE2+pIdrXvJj/nmY2xMwyzSyzZs2acZ7CVTSK+Kva\nid+4j7uZxyEcyXRuYACt+Ywsjij0OJ4sXEUWzyCH1wP9Cb7trySYyOmGwnYys+LeJLgS8vR3rAes\nLuYxXQWlfF9DTucDBtCNg/iK4VzKbTxe4Ex8kTxhuIquwMQRNlBfbmadSimeSDOARpIOAFYBlwCX\npiAOV85FJo06rOJJbqYDb7CYgzmZ8Uzg5EKP4cnCuT8VWFVlZltJQoO0pPMkrQSOAkZL+jBcX0fS\nmPDcW4BuwIcEVWOvmdmCko7Fpaf87RiV2cLNPEE2TTiL97iTB2jBXE8azhWBrJD/Ckl9gBrACGBj\n7noz+zy5oRVdZmamZWVlpToMlwL5q6QA2jCNwXShFXMYQ1u6MaDQ6Vs9WbiKRtJMM4t5X12keNo4\njg4f749YZ8DfEg3MuWSJljD25Cf60ovODGUldbmAN3iL8/HpW50rnkITh5mdVBqBOFcU0RIGGFcy\njEf5B3uyjse5lXu5l1/ZLeZxPFk4F7947hzfV9Jzkt4Pl5tKuib5oTlXsGhJoykLmMQJvMDf+ZJG\nHMbn9OBxTxrOlaB47uN4gaCBuk64/AVwc7ICci4e+ZNGdTbSl57MphXNWMA1/Jtj+aTA6Vv9Xgzn\niiaexLGPmb0GbIPtvZ1K485x5/LI31MqV3tGspCm9OQRXuQKGrOY/3ANVsCftycM54ounsbxjZL2\nJrxrW1Ib4JekRuVchOjtGNCAZTxFd9rzLvNozrH8jykcG/M4niycKxnxJI5bgVHAXyVNIZiP48Kk\nRuUcsRNGVf7gNh7nLh5gG5XowaP056aYM/F5wnCuZMXTq+pzSScAjQn6MS42s81Jj8xVSLGSRa4T\nmMggutKURbzFedxE/5gz8XnCcC45YiYOSefHeOlgSZjZW0mKyVVAhSWMmnzPY/TgCl5iKQ05k/cY\nw5lRt/WE4VxyFVTiODt8rEVwE+CEcPkkYCLgicMVS2HJAkBsozNDeJg72IWN9OGf9KF3zJn4PGk4\nl3wxE4eZ/R1A0ntAUzNbEy7XJpiZz7mExZMscrViFs9wPa2ZzgRO4gYGkk1GzO09aThXOuLpjtsw\nN2mEvgMOTlI8Lg3F6kYby26s50luIotMGrJs+/St0ZKGT6TkXOmLp1fVxHD02v8SdMm9BPg4qVG5\nci+RksWfjIt4jSe4hf34lsF04U4e5Gf23HFLTxTOpUw8vaq6STqPP+cZH2Jmbyc3LFdeFS1hwEF8\nyUBu4DTGMZPDOIeRUWfi84ThXOrFM5HTh+Fsfp4sXFRFTRYAf+F3etGXXvRlE3+hG08zmC5so/L2\nbTxZOFe2FJg4zGyrpBxJNczM7xZ3eRQnYQCcylgGcgONWMIrdOQ2Hudbam9/3ROGc2VTPG0cvwPz\nJI0j70RO3ZMWlSvTipsw6rCKftzKxbzGYg7mFMbxEcEU9Z4snCv74kkco8MfV8EVN2FUZgvdGMD9\n3E01/uAu7ucRbmeT/aVkAnTOlYp4uuOOAGYCWcAIMxtmZsOSG5YrS4YPL37SaM2nzOAInuQWpnAM\nOy1ZwAN2lycN58qhgoYcqQJfE1W6AAAVVElEQVQ8BFwNfEOQZOpJeh7o7eNVpb/iJgsIpm99mDu4\njqFUqlsHnnydthdcUDIHd86lREEljkeBvYADzOxwMzsU+CuwB/BYaQTnSl+iN+vFZlzBMLJpwv9V\nfo5Kt94CixbBhRd60nCunCuojeMs4GCzP5srzWy9pC5ANnBTsoNzpaOkP8ebsoAFx3eFyZPhqKNg\n8GBo2bJkT+KcS5mCShwWmTQiVm4lnNTJlX8lmTSqsxHr2YsFVVrB/PkwdCh88oknDefSTEGJY6Gk\nK/KvlHQZQYmjyCR1kLRA0jZJmQVst0zSPEmzJWUV55xuRyWZNGzkKDbWbwr/+hdcfjlkZ8O110Kl\nePpfOOfKk4Kqqm4A3pJ0NUGvKgOOAHYGzivmeecD5wPPxrHtSWb2QzHP5/IpqaRhy76B7t3hnFHQ\nrFlQPXXccSVzcOdcmVTQsOqrgNaS/gY0I5j9730z+6i4JzWzRQDyRtKUKInLbpv+gCeegIz7ggM+\n8gjcfDNUjT59q3MufcQzyOEE/pzEqbQZMFaSAc+a2ZBYG0rqDHQGqF+/fimFV/4UN2mYAZMmwaFd\nYeFCOPdc6N8f/Jo7V2HEc+d4kUgaD+wX5aXeZjYyzsMcY2arJdUCxknKNrPJ0TYMk8oQgMzMTG+8\nj6KoSWN7F4nvv4cr/wEvvggNG8K778JZZ5VUeM65ciJpiSMcUbe4x1gdPn4v6W3gSCBq4nAFSzRp\n5OlPt21b0EPqjjvg11+DxzvvhOrRp291zqW3MtvlRdIuknbLfQ6cRtCo7hKU6Od7nqQxaxYcfTRc\nf33QrXbOHHjoIU8azlVgKUkcks6TtBI4ChgdzjCIpDqSxoSb7Qt8ImkOMB0YbWYfpCLe8u633+Lf\ndnvSWL8+aOzOzISlS+Gll2DCBMiIPee3c65iSFpVVUHCGQR3mBgqrJpqFz7/GvA7x4qpWrX4t+3S\nhSBzvP463HILrFkTlDT69IE9d5y+1TlXMaUkcbjSsznOoSi7dIFBty6BM26AsWPhsMPg7bfhyCOT\nG6BzrtzxxJHG4m0Qt99+D+74bv5wUER56ino2hUqVy58Z+dcheOJI03FnTTGjoNDusKSJXDJJdCv\nH9SuXfiOzrkKq8z2qnLJVZvVvF75YjjttCDLjB0L//2vJw3nXKE8cVQwldlCd/qTTRMurDIS7rsP\n5s6FU09NdWjOuXLCq6oqkCP5jMF04TBm8QGnc8b8AXDQQakOyzlXzniJIw0NH553eQ/WMZjrmcZR\n7Mt3dOA1ztj2vicN51yReOJIQ9dem/vMuJwXWUxjrmMo/bmJDBbxunXw6Vudc0XmVVVp6PffIYOF\nDKIrJzKJabThNMYyh1apDs05lwa8xJFucnJ4iDuYQ0taMJfrGMIxTPGk4ZwrMZ440sm770LTptxB\nX17mMhqzmH9zHRbxa3755RTG55xLC5440sHy5cGESu3bw667chyTuZrn+YGaO2zaqVMK4nPOpRVP\nHOXZ5s3BlK0ZGTBuXDBsyKxZfILP+e2cSx5vHC+vJk8ORiZcuBDOOSeYvrVBg1RH5ZyrALzEUd6s\nXQtXXQUnnAAbN8KoUfDOO540nHOlxhNHebFtGwwZAo0bB3f49eoFCxbA2WenOjLnXAXjVVXlwezZ\nQbXUp58GJY1Bg6Bp05ibS/mmf41Y75xzxeUljrJsw4ZgJr7DD4evvoJhw+DjjwtMGhBM2pfIeuec\nS4SXOMoiM3jjjWDO7zVroHNnePjhuKdvHTQoeBwyBLZuDeZj6tz5z/XOOVccXuIoa5YsgbZt4aKL\noFYtmDYNnnkm4Tm/jzkG6tULqqfq1QuWnXOuJHiJo6zYtCm4D+Ohh4LpW/v3D6ZvrZL4r2j48KCE\nkZMTLH/zTbAMfgOgc674vMRRFowfD4ccAvfcE9wBnp0N3bsXKWkA9O79Z9LIlZMTrHfOueLyxJFK\na9ZAx47B7Htm8OGH8OqrUKdOsQ67fHli651zLhEpSRySHpWULWmupLcl7RFjuzMkLZa0RFKv0o4z\nabZuhaefhiZN4O234d57Yd68YP7vElC/fmLrnXMuEakqcYwDmptZC+AL4I78G0iqDAwE2gJNgY6S\nCu6HWh5Mnw5HHBFURbVpEySMe+6BnXYqsVP06QPVq+ddV716sN4554orJYnDzMaa2ZZw8VOgXpTN\njgSWmNnXZvYH8CpwTmnFWOLWrQtu4mvTBr79FkaMgA8+gEaNSvxUnToFXXEbNAh6VTVoECx7w7hz\nriSUhV5VVwMjoqyvC6yIWF4JtI51EEmdgc4A9ctSnYxZMAlGjx7www9BSeP++2H33ZN62k6dPFE4\n55IjaYlD0nhgvygv9TazkeE2vYEtwPBoh4iyLspAGuELZkOAIQCZmZkxtytVixYFXWonToTWrYMS\nxqGHpjoq55wrlqQlDjM7paDXJV0JnAWcbBZtZCVWAvtHLNcDVpdchEmUkwMPPgiPPQa77ALPPgvX\nXguVvBObc678S0lVlaQzgJ7ACWaWE2OzGUAjSQcAq4BLgEtLKcSie+89uPFGWLYMrrwymGipVq1U\nR+WccyUmVV+BBwC7AeMkzZb0DICkOpLGAISN592AD4FFwGtmtiBF8RZu+XI477xgmPPq1WHSJHjh\nBU8azrm0k5ISh5kdFGP9aqBdxPIYYExpxVUkmzfDk08G92KYQd++wYi21aqlOjLnnEuKstCrqvz6\n5JOgi+38+dC+PTz1lM/E55xLe95aWxRr18LVV8Nxx8H69TByZPDjScM5VwF44kjEtm0wdGgwVMhL\nL0HPnrBwYVDacM65CsKrquI1Z05QLTVtGhx/fDArUrNmqY7KOedKnZc4CrNhA9x6azB965dfBtO3\nTpzoScM5V2F5iSMWM3jzzWD61tWrg5mQHnoI9tor1ZE551xKeYkjmq++gnbtoEMHqFkTpk4Npm/1\npOGcc5448ti0CR54AJo3hylTgvszZswIRrR1zjkHeFXVn9atCxLEF1/ARRdBv35Qt26qo3LOuTLH\nSxy59twTzjwzmL51xAhPGs6VknfeeYfrrruOc845h7Fjx6Y6HBcHTxyR+vUrselbnXN5Pfvss+y3\n3360bNmSv/71r7z44osAnHvuuQwdOpQXXniBESOiTc0Tnw8++IDGjRtz0EEH0bdv35jb9e/fn+bN\nm9OsWTOefPLJ7eufeOIJmjVrRvPmzenYsSO///779teuvvpqatWqRfPmzbevW7FiBSeddBIZGRk0\na9aM/v37Fzn2csfM0u7n8MMPN+dc2dK1a1cbPHiwmZl99tlntvfee+d5/dZbb7WZM2cW6dhbtmyx\nAw880L766ivbtGmTtWjRwhYsWLDDdvPmzbNmzZrZxo0bbfPmzXbyySfbF198YStXrrSGDRtaTk6O\nmZl16NDBnn/++e37TZo0yWbOnGnNmjXbvm716tXb412/fr01atQo6jnLCyDL4vyM9RKHc65UzJs3\nj8aNGwNwwAEHUC0cCNTM6NmzJ23btuWwww4r0rGnT5/OQQcdxIEHHki1atW45JJLGDly5A7bLVq0\niDZt2lC9enWqVKnCCSecwNtvvw3Ali1b+O2339iyZQs5OTnUqVNn+37HH388e+XrVVm7du3t8e62\n225kZGSwatWqIsVf3njicM6VitzEYWYMGDCAPn36APD0008zfvx43njjDZ555pkd9jvuuONo1arV\nDj/jx4/fvs2qVavYf/8/532rV69e1A/x5s2bM3nyZH788UdycnIYM2YMK1asoG7duvTo0YP69etT\nu3ZtatSowWkJVFsvW7aMWbNm0bp1zNmt04r3qnLOJd2KFSvYsGED7dq1Y9WqVbRo0YJ7770XgO7d\nu9O9e/eY+/7vf/8r9PgWZRJRacfZpzMyMujZsyennnoqu+66Ky1btqRKlSqsW7eOkSNHsnTpUvbY\nYw86dOjAyy+/zGWXXVbouX/99VcuuOACnnzySXbfffdCt08HXuJwziXd3LlzOf7445k9ezZffPEF\n2dnZTJs2La594ylx1KtXjxUrVmxfXrlyZZ6qpkjXXHMNn3/+OZMnT2avvfaiUaNGjB8/ngMOOICa\nNWtStWpVzj//fKZOnVpobJs3b+aCCy6gU6dOnH/++XG9n3TgJQ7nXNLNmzePQw89FIA999yTSy+9\nlNGjR3P00UcXum88JY4jjjiCL7/8kqVLl1K3bl1effVVXnnllajbfv/999SqVYvly5fz1ltvMW3a\nNL744gs+/fRTcnJy2Hnnnfnoo4/IzMws8JxmxjXXXENGRga33nproTGmEy9xhIYPh4YNoVKl4HH4\n8Ojru3YNHiWoUiXvY6VKwWNZ/XEuVSITB8DZZ5/NmDElN7lnlSpVGDBgAKeffjoZGRlcdNFFNIsY\niLRdu3asXr0agAsuuICmTZty9tlnM3DgQPbcc09at27NhRdeyGGHHcYhhxzCtm3b6Ny58/b9O3bs\nyFFHHcXixYupV68ezz33HFOmTOGll15iwoQJ20tBJfmeyjJFqxss7zIzMy0rKyvu7YcPD8YwzMn5\nc1316nDllcFguJHry7s0/HU750qApJlmVnAxK+RVVUDv3jsmh5wcGDIEtm5NTUzOOVdWeVUVsHx5\n9PWeNJxzbkeeOID69aOvr1y5dONwzrnyICWJQ9KjkrIlzZX0tqQ9Ymy3TNI8SbMlxd9okaA+fYI2\njUjVqwftHvnXO+dcRZeqEsc4oLmZtQC+AO4oYNuTzKxVvI02RdGpU9Ce0aBB0PuoQYNgedCgHdd3\n6RI8wp8lktzHst5zyRvGnXMlIeW9qiSdB1xoZp2ivLYMyDSzHxI5ZqK9qpxzrqJLpFdVWWjjuBp4\nP8ZrBoyVNFNS5xjbOOecK0VJ644raTywX5SXepvZyHCb3sAWYHiMwxxjZqsl1QLGSco2s8kxztcZ\n6AxQP1Zrt3POuWJLWuIws1MKel3SlcBZwMkWo77MzFaHj99Lehs4EoiaOMxsCDAEgqqqYoTunHOu\nAKnqVXUG0BNob2ZR78uWtIuk3XKfA6cB80svSuecc9Gkqo1jALAbQfXTbEnPAEiqIyl3sJd9gU8k\nzQGmA6PN7IPUhOuccy5XyntVJYOktcA3KTr9PkBCvcDSnF+PvPx65OXXI69UXo8GZlYzng3TMnGk\nkqSsZN5zUt749cjLr0defj3yKi/Xoyx0x3XOOVeOeOJwzjmXEE8cJW9IqgMoY/x65OXXIy+/HnmV\ni+vhbRzOOecS4iUO55xzCfHE4ZxzLiGeOJJIUg9JJmmfVMeSSvHOv5LOJJ0habGkJZJ6pTqeVJK0\nv6SPJS2StEDSTamOqSyQVFnSLEnvpTqWwnjiSBJJ+wOnAjEmpq1QEpl/Je1IqgwMBNoCTYGOkpqm\nNqqU2gLcZmYZQBvghgp+PXLdBCxKdRDx8MSRPE8AtxMMDV+hmdlYM9sSLn4K1EtlPClwJLDEzL42\nsz+AV4FzUhxTypjZGjP7PHy+geDDsm5qo0otSfWAM4F/pzqWeHjiSAJJ7YFVZjYn1bGUQQXNv5Ku\n6gIrIpZXUsE/KHNJaggcCnyW2khS7kmCL5rbUh1IPJI2rHq6K2i+EeCfBKP5VhglNP9Kuoo2qXCF\nL4lK2hV4E7jZzNanOp5UkXQW8L2ZzZR0YqrjiYcnjiKKNd+IpEOAA4A5CiYhrwd8LulIM/u2FEMs\nVSUx/0oaWwnsH7FcD1idoljKBElVCZLGcDN7K9XxpNgxQHtJ7YCdgN0lvWxml6U4rpj8BsAkK+q8\n6ekknH+lH3CCma1NdTylTVIVgk4BJwOrgBnApWa2IKWBpYiCb1TDgJ/M7OZUx1OWhCWOHmZ2Vqpj\nKYi3cbjSEHX+lYoi7BjQDfiQoCH4tYqaNELHAJcDfwv/HmaH37ZdOeElDueccwnxEodzzrmEeOJw\nzjmXEE8czjnnEuKJwznnXEI8cTjnnEuIJw5XZknaO6K75reSVoXPf5a0sJRjaRXZZVRS+6KOcitp\nWapGTJZ0laQ6Ecv/zh1gMJVxufLFE4crs8zsRzNrZWatgGeAJ8LnrUjCmD7hjXqxtAK2Jw4zG2Vm\nfUs6hlJwFbA9cZjZtWZWqknYlX+eOFx5VVnS0HA+h7GSdgaQ9FdJH0iaKel/kpqE6xtI+iicE+Qj\nSfXD9S9I6ifpY+BfknaR9B9JM8K5Ec6RVA24H7g4LPFcHH5zHxAeY99wnpE54c/R4fp3wjgWSOpc\n2BuS9HdJX0iaFL633OO/IOnCiO1+DR93Dd/L55LmSTonXN8wnOsiz/UJj5EJDA/fx86SJkrKjBLL\nZZKmh9s9G84VUTmMZX54vluK8ftz5ZgnDldeNQIGmlkz4GfggnD9EOBGMzsc6AEMCtcPAF4M5wQZ\nDjwVcayDgVPM7DaCQSonmNkRwEnAo0BV4G5gRFgCGpEvlqeASWbWEjgMyL0r/Oowjkygu6S9Y70Z\nSbWB+wjuqj6VYN6OwvwOnGdmh4WxPh4O5xH1+pjZG0AW0Cl8H7/FiCUDuBg4JizhbQU6EZS66ppZ\nczM7BHg+jhhdGvJBDl15tdTMZofPZwINw9FWjwZe//Pzk7+Ej0cB54fPXwIeiTjW62a2NXx+GsGA\ncz3C5Z2A+oXE8jfgCoDwOL+E67tLOi98vj/Bh/mPMY7RGpiYO5aXpBEECa0gAh6SdDxB1V1dYN/w\ntR2uTyHHinQycDgwI7yOOwPfA+8CB0p6GhgNjE3gmC6NeOJw5dWmiOdbCT7cKgE/h9+SCxM51s7G\niOci+Ha+OHJjSa0TCS4crO4U4Cgzy5E0kSAJxRtTpC2EtQNhiaJauL4TUBM43Mw2hwNq5p4j2vWJ\nO3xgmJntMFOjpJbA6cANwEUE86u4CsarqlzaCOd0WCqpAwQfsuEHHcBU4JLweSfgkxiH+RC4MbfK\nR9Kh4foNBAM1RvMR0CXcvrKk3YEawLowaTQhmCK1IJ8BJ4Y9yaoCHSJeW0ZQAoBg5sCq4fMaBPM4\nbJZ0EtCgkHMU9j4i38+FkmqF72mvsI1oH6CSmb0J3EVQLecqIE8cLt10Aq6RNIegrSF3itbuwN8l\nzSUYmfWmGPs/QPDBPFfS/HAZ4GOgaW7jeL59bgJOkjSPoFqoGfABUCU83wMEU+bGZGZrgHuBacB4\n4POIl4cCJ0iaTlCllVtCGg5kSsoK33d2QecIvQA8k9s4HiOWhcCdwNgw/nFAbYKqsImSZofHqVBz\nx7s/+ei4zpVBkq4imMelW6pjcS4/L3E455xLiJc4nHPOJcRLHM455xLiicM551xCPHE455xLiCcO\n55xzCfHE4ZxzLiH/DzIKspdCbvk8AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2730604be10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# show difference from normal distribution of transformed feature\n",
    "# Code adapated from here: https://www.kaggle.com/snmateen/simple-eda-feature-transformations/notebook\n",
    "\n",
    "ax=plt.gca()\n",
    "stats.probplot(xcf, dist=stats.norm, plot=ax, rvalue=True)\n",
    "ax.set_title('Probplot of transformed feature')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAD8CAYAAACMwORRAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ncej/A6jprTH9WkxxAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x27306568f98>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Code adapated from here: https://www.kaggle.com/snmateen/simple-eda-feature-transformations/notebook\n",
    "\n",
    "sns.distplot(xcf, bins=50)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "metadata": {},
   "outputs": [],
   "source": [
    "# make a new df with the transformed feature\n",
    "train.cont7 = xcf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>cont3</th>\n",
       "      <th>cont7</th>\n",
       "      <th>cont4</th>\n",
       "      <th>cont2</th>\n",
       "      <th>cont5</th>\n",
       "      <th>cont14</th>\n",
       "      <th>cont8</th>\n",
       "      <th>cont13</th>\n",
       "      <th>cont6</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>185659.000000</td>\n",
       "      <td>185659.000000</td>\n",
       "      <td>185659.000000</td>\n",
       "      <td>185659.000000</td>\n",
       "      <td>185659.000000</td>\n",
       "      <td>185659.000000</td>\n",
       "      <td>185659.000000</td>\n",
       "      <td>185659.000000</td>\n",
       "      <td>185659.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>0.497367</td>\n",
       "      <td>-0.771165</td>\n",
       "      <td>0.494039</td>\n",
       "      <td>0.506920</td>\n",
       "      <td>0.489335</td>\n",
       "      <td>0.495556</td>\n",
       "      <td>0.487035</td>\n",
       "      <td>0.493952</td>\n",
       "      <td>0.487960</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.201635</td>\n",
       "      <td>0.330414</td>\n",
       "      <td>0.211590</td>\n",
       "      <td>0.207063</td>\n",
       "      <td>0.209507</td>\n",
       "      <td>0.222310</td>\n",
       "      <td>0.200028</td>\n",
       "      <td>0.212223</td>\n",
       "      <td>0.204323</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.002634</td>\n",
       "      <td>-2.411702</td>\n",
       "      <td>0.176921</td>\n",
       "      <td>0.001149</td>\n",
       "      <td>0.281143</td>\n",
       "      <td>0.179722</td>\n",
       "      <td>0.236880</td>\n",
       "      <td>0.000228</td>\n",
       "      <td>0.012683</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>0.336963</td>\n",
       "      <td>-1.008256</td>\n",
       "      <td>0.327354</td>\n",
       "      <td>0.358319</td>\n",
       "      <td>0.281143</td>\n",
       "      <td>0.294644</td>\n",
       "      <td>0.312800</td>\n",
       "      <td>0.315758</td>\n",
       "      <td>0.334532</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>0.527991</td>\n",
       "      <td>-0.805101</td>\n",
       "      <td>0.452887</td>\n",
       "      <td>0.555782</td>\n",
       "      <td>0.422268</td>\n",
       "      <td>0.408085</td>\n",
       "      <td>0.441060</td>\n",
       "      <td>0.363547</td>\n",
       "      <td>0.438917</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>0.634224</td>\n",
       "      <td>-0.522034</td>\n",
       "      <td>0.670373</td>\n",
       "      <td>0.681761</td>\n",
       "      <td>0.643315</td>\n",
       "      <td>0.724558</td>\n",
       "      <td>0.623580</td>\n",
       "      <td>0.689974</td>\n",
       "      <td>0.648884</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>0.944251</td>\n",
       "      <td>-0.049047</td>\n",
       "      <td>0.954297</td>\n",
       "      <td>0.862654</td>\n",
       "      <td>0.983674</td>\n",
       "      <td>0.844848</td>\n",
       "      <td>0.980200</td>\n",
       "      <td>0.988494</td>\n",
       "      <td>0.997162</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               cont3          cont7          cont4          cont2  \\\n",
       "count  185659.000000  185659.000000  185659.000000  185659.000000   \n",
       "mean        0.497367      -0.771165       0.494039       0.506920   \n",
       "std         0.201635       0.330414       0.211590       0.207063   \n",
       "min         0.002634      -2.411702       0.176921       0.001149   \n",
       "25%         0.336963      -1.008256       0.327354       0.358319   \n",
       "50%         0.527991      -0.805101       0.452887       0.555782   \n",
       "75%         0.634224      -0.522034       0.670373       0.681761   \n",
       "max         0.944251      -0.049047       0.954297       0.862654   \n",
       "\n",
       "               cont5         cont14          cont8         cont13  \\\n",
       "count  185659.000000  185659.000000  185659.000000  185659.000000   \n",
       "mean        0.489335       0.495556       0.487035       0.493952   \n",
       "std         0.209507       0.222310       0.200028       0.212223   \n",
       "min         0.281143       0.179722       0.236880       0.000228   \n",
       "25%         0.281143       0.294644       0.312800       0.315758   \n",
       "50%         0.422268       0.408085       0.441060       0.363547   \n",
       "75%         0.643315       0.724558       0.623580       0.689974   \n",
       "max         0.983674       0.844848       0.980200       0.988494   \n",
       "\n",
       "               cont6  \n",
       "count  185659.000000  \n",
       "mean        0.487960  \n",
       "std         0.204323  \n",
       "min         0.012683  \n",
       "25%         0.334532  \n",
       "50%         0.438917  \n",
       "75%         0.648884  \n",
       "max         0.997162  "
      ]
     },
     "execution_count": 122,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train[cols_cont_feat].describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "metadata": {},
   "outputs": [],
   "source": [
    "# split data into train and test dfs\n",
    "X_train, X_test, y_train, y_test = train_test_split(train, train_loss, test_size=.2, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0:01:05.152434\n",
      "The mae of the model prediction is:  1126.7991885\n",
      "\n",
      "Best params:\n",
      "{'learning_rate': 0.2, 'min_data_in_leaf': 375, 'n_estimators': 500, 'num_leaves': 12}\n"
     ]
    }
   ],
   "source": [
    "# Final parameters\n",
    "\n",
    "param_grid = {\n",
    "    'learning_rate': [.2],\n",
    "    'n_estimators': [500],\n",
    "    'num_leaves': [12],\n",
    "    'min_data_in_leaf': [375]\n",
    "}\n",
    "\n",
    "start=datetime.now()\n",
    "\n",
    "lgbm_reg_best = GridSearchCV(lgbm_reg, param_grid)\n",
    "lgbm_reg_best.fit(X_train, y_train)\n",
    "\n",
    "stop=datetime.now()\n",
    "print(stop-start)\n",
    "\n",
    "y_pred=lgbm_reg_best.predict(X_test)\n",
    "\n",
    "adj_y_test = np.exp(y_test) - shift\n",
    "adj_y_pred = np.exp(y_pred) - shift\n",
    "\n",
    "#print('The mae of the model prediction is: ', mean_absolute_error(y_test, y_pred))\n",
    "print('The mae of the model prediction is: ', mean_absolute_error(adj_y_test, adj_y_pred))\n",
    "\n",
    "print('\\nBest params:\\n{}'.format(lgbm_reg_best.best_params_))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "metadata": {},
   "outputs": [],
   "source": [
    "# *****\n",
    "# This was the final model (tried PCA below but it didn't help)\n",
    "# The mae of the model prediction was:  1126.7991885\n",
    "# *****"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0:00:00.093246\n"
     ]
    }
   ],
   "source": [
    "# Try PCA on remaining continuous features\n",
    "# This example was helpful: https://www.kaggle.com/bertcarremans/reducing-number-of-numerical-features-with-pca\n",
    "\n",
    "\n",
    "# scale cont7 (boxcox changed scale - it was 0-1 like the other cont features)\n",
    "# Removing because cont. features are already scaled\n",
    "from sklearn.preprocessing import minmax_scale\n",
    "\n",
    "minmax_scale(xcf, feature_range=(0,1), copy=False)\n",
    "\n",
    "pca = PCA(n_components=None, random_state=42)\n",
    "\n",
    "start=datetime.now()\n",
    "train_pca = pca.fit_transform(train[cols_cont_feat])\n",
    "stop=datetime.now()\n",
    "print(stop-start)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfUAAAF3CAYAAAC8MNLCAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x273057e85c0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# graph showing amount of variance explained \n",
    "# code adapted from here: https://www.kaggle.com/bertcarremans/reducing-number-of-numerical-features-with-pca \n",
    "\n",
    "varExp = pca.explained_variance_ratio_\n",
    "\n",
    "# based on example from here: https://stackoverflow.com/questions/15889131/how-to-find-the-cumulative-sum-of-numbers-in-a-list \n",
    "vey = np.cumsum(varExp)\n",
    "oy = [x for x in range(1, len(varExp)+1)]\n",
    "\n",
    "plt.subplots(figsize=(8, 6))\n",
    "plt.grid(True)\n",
    "plt.plot(oy, vey, 'bo-')\n",
    "plt.ylabel('Cumulative Explained Variance')\n",
    "plt.xlabel('Number of Components')\n",
    "plt.ylim([0.0, 1.1])\n",
    "plt.xticks(oy)\n",
    "plt.yticks(np.arange(0.0, 1.1, 0.10))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0:00:00.365974\n"
     ]
    }
   ],
   "source": [
    "# Run PCA on train dataset with 7 components, which explain 95% of variance\n",
    "pca = PCA(n_components=7, random_state=42)\n",
    "\n",
    "start=datetime.now()\n",
    "train_pca_cont_feat = pd.DataFrame(pca.fit_transform(train[cols_cont_feat]), columns=['pca_cont1','pca_cont2','pca_cont3','pca_cont4','pca_cont5','pca_cont6','pca_cont7'])\n",
    "stop=datetime.now()\n",
    "print(stop-start)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>pca_cont1</th>\n",
       "      <th>pca_cont2</th>\n",
       "      <th>pca_cont3</th>\n",
       "      <th>pca_cont4</th>\n",
       "      <th>pca_cont5</th>\n",
       "      <th>pca_cont6</th>\n",
       "      <th>pca_cont7</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>1.856590e+05</td>\n",
       "      <td>1.856590e+05</td>\n",
       "      <td>1.856590e+05</td>\n",
       "      <td>1.856590e+05</td>\n",
       "      <td>1.856590e+05</td>\n",
       "      <td>1.856590e+05</td>\n",
       "      <td>1.856590e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>7.563682e-17</td>\n",
       "      <td>1.708279e-17</td>\n",
       "      <td>-1.190406e-16</td>\n",
       "      <td>4.173913e-17</td>\n",
       "      <td>8.343222e-17</td>\n",
       "      <td>-7.840336e-17</td>\n",
       "      <td>1.635443e-16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>4.000464e-01</td>\n",
       "      <td>2.930905e-01</td>\n",
       "      <td>2.517791e-01</td>\n",
       "      <td>2.212319e-01</td>\n",
       "      <td>1.876488e-01</td>\n",
       "      <td>1.797001e-01</td>\n",
       "      <td>1.359113e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>-1.069253e+00</td>\n",
       "      <td>-9.549809e-01</td>\n",
       "      <td>-7.993702e-01</td>\n",
       "      <td>-4.208201e-01</td>\n",
       "      <td>-5.476038e-01</td>\n",
       "      <td>-7.793232e-01</td>\n",
       "      <td>-5.329009e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>-3.059459e-01</td>\n",
       "      <td>-1.742150e-01</td>\n",
       "      <td>-1.827656e-01</td>\n",
       "      <td>-1.989462e-01</td>\n",
       "      <td>-1.497158e-01</td>\n",
       "      <td>-1.163898e-01</td>\n",
       "      <td>-9.210513e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>1.876178e-02</td>\n",
       "      <td>1.308574e-02</td>\n",
       "      <td>-2.690152e-03</td>\n",
       "      <td>-7.029353e-02</td>\n",
       "      <td>6.592054e-03</td>\n",
       "      <td>8.509964e-03</td>\n",
       "      <td>-2.387026e-03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>3.318876e-01</td>\n",
       "      <td>2.292985e-01</td>\n",
       "      <td>1.820982e-01</td>\n",
       "      <td>2.267421e-01</td>\n",
       "      <td>1.425138e-01</td>\n",
       "      <td>1.211884e-01</td>\n",
       "      <td>8.702271e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>1.723468e+00</td>\n",
       "      <td>1.120669e+00</td>\n",
       "      <td>8.100631e-01</td>\n",
       "      <td>4.422685e-01</td>\n",
       "      <td>8.528357e-01</td>\n",
       "      <td>9.057689e-01</td>\n",
       "      <td>5.700605e-01</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          pca_cont1     pca_cont2     pca_cont3     pca_cont4     pca_cont5  \\\n",
       "count  1.856590e+05  1.856590e+05  1.856590e+05  1.856590e+05  1.856590e+05   \n",
       "mean   7.563682e-17  1.708279e-17 -1.190406e-16  4.173913e-17  8.343222e-17   \n",
       "std    4.000464e-01  2.930905e-01  2.517791e-01  2.212319e-01  1.876488e-01   \n",
       "min   -1.069253e+00 -9.549809e-01 -7.993702e-01 -4.208201e-01 -5.476038e-01   \n",
       "25%   -3.059459e-01 -1.742150e-01 -1.827656e-01 -1.989462e-01 -1.497158e-01   \n",
       "50%    1.876178e-02  1.308574e-02 -2.690152e-03 -7.029353e-02  6.592054e-03   \n",
       "75%    3.318876e-01  2.292985e-01  1.820982e-01  2.267421e-01  1.425138e-01   \n",
       "max    1.723468e+00  1.120669e+00  8.100631e-01  4.422685e-01  8.528357e-01   \n",
       "\n",
       "          pca_cont6     pca_cont7  \n",
       "count  1.856590e+05  1.856590e+05  \n",
       "mean  -7.840336e-17  1.635443e-16  \n",
       "std    1.797001e-01  1.359113e-01  \n",
       "min   -7.793232e-01 -5.329009e-01  \n",
       "25%   -1.163898e-01 -9.210513e-02  \n",
       "50%    8.509964e-03 -2.387026e-03  \n",
       "75%    1.211884e-01  8.702271e-02  \n",
       "max    9.057689e-01  5.700605e-01  "
      ]
     },
     "execution_count": 129,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_pca_cont_feat.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_pca_cont_feat.reset_index(drop=True, inplace=True)\n",
    "train.reset_index(drop=True, inplace=True)\n",
    "train_pca = pd.concat([train_pca_cont_feat, train[train.columns.difference(cols_cont_feat)]], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "185659 185659\n"
     ]
    }
   ],
   "source": [
    "# confirm shape of new df\n",
    "print(train_pca.shape[0], train.shape[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 132,
   "metadata": {},
   "outputs": [],
   "source": [
    "# split data into train and test dfs\n",
    "X_train, X_test, y_train, y_test = train_test_split(train_pca, train_loss, test_size=.2, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0:01:18.619059\n",
      "The mae of the model prediction is:  1134.33426738\n",
      "\n",
      "Best params:\n",
      "{'learning_rate': 0.2, 'min_data_in_leaf': 375, 'n_estimators': 500, 'num_leaves': 12}\n"
     ]
    }
   ],
   "source": [
    "lgbm_reg_pca = lgb.LGBMRegressor(random_state=42)\n",
    "\n",
    "# Final parameters\n",
    "param_grid = {\n",
    "    'learning_rate': [.2],\n",
    "    'n_estimators': [500],\n",
    "    'num_leaves': [12],\n",
    "    'min_data_in_leaf': [375]\n",
    "}\n",
    "\n",
    "start=datetime.now()\n",
    "\n",
    "lgbm_reg_best_pca = GridSearchCV(lgbm_reg_pca, param_grid)\n",
    "lgbm_reg_best_pca.fit(X_train, y_train)\n",
    "\n",
    "stop=datetime.now()\n",
    "print(stop-start)\n",
    "\n",
    "y_pred=lgbm_reg_best_pca.predict(X_test)\n",
    "\n",
    "# adjust the predictions (due to adjustment of loss)\n",
    "adj_y_test = np.exp(y_test) - shift\n",
    "adj_y_pred = np.exp(y_pred) - shift\n",
    "\n",
    "#print('The mae of the model prediction is: ', mean_absolute_error(y_test, y_pred))\n",
    "print('The mae of the model prediction is: ', mean_absolute_error(adj_y_test, adj_y_pred))\n",
    "\n",
    "print('\\nBest params:\\n{}'.format(lgbm_reg_best_pca.best_params_))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "metadata": {},
   "outputs": [],
   "source": [
    "# \n",
    "# No improvement from PCA\n",
    "#"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 136,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Preprocess validation features (from test.csv file) to conform with sucessful feature preprocessing of train data:\n",
    "\n",
    "#   Drop features highly correlated with cont6\n",
    "validate = validate[validate.columns.difference(cols_cont_drop)]\n",
    "\n",
    "#   Box-cox transform cont7 feature\n",
    "validate.cont7, _ = stats.boxcox(validate.cont7)\n",
    "\n",
    "#   Run model on processed validation data\n",
    "y_pred = lgbm_reg_best.predict(validate)\n",
    "\n",
    "#   Transform loss predictions back to original feature space:\n",
    "adj_y_pred = np.exp(y_pred) - shift"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 137,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>loss</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>125546.000000</td>\n",
       "      <td>125546.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>294067.153442</td>\n",
       "      <td>2668.323054</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>170098.335649</td>\n",
       "      <td>1957.193377</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>4.000000</td>\n",
       "      <td>357.507791</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>146414.250000</td>\n",
       "      <td>1408.635515</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>294306.000000</td>\n",
       "      <td>2056.830177</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>441800.000000</td>\n",
       "      <td>3218.161413</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>587634.000000</td>\n",
       "      <td>25161.196432</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  id           loss\n",
       "count  125546.000000  125546.000000\n",
       "mean   294067.153442    2668.323054\n",
       "std    170098.335649    1957.193377\n",
       "min         4.000000     357.507791\n",
       "25%    146414.250000    1408.635515\n",
       "50%    294306.000000    2056.830177\n",
       "75%    441800.000000    3218.161413\n",
       "max    587634.000000   25161.196432"
      ]
     },
     "execution_count": 137,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "val_pred = pd.DataFrame({'id': validate_orig.id, 'loss': adj_y_pred})\n",
    "val_pred.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "val_pred.to_csv(\".\\output\\submit2.csv\", index=False, float_format=\"%.2f\")\n",
    "# score from Kaggle: 1131.90796\n",
    "# rank: between entries 1350 and 1351"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0:00:42.683519\n"
     ]
    }
   ],
   "source": [
    "# this code is based on the example in the sklearn documentation\n",
    "\n",
    "from sklearn.neighbors import KNeighborsRegressor\n",
    "\n",
    "knnr = KNeighborsRegressor()\n",
    "\n",
    "start=datetime.now()\n",
    "knnr.fit(X_train, y_train)\n",
    "stop=datetime.now()\n",
    "print(stop-start)\n",
    "\n",
    "start=datetime.now()\n",
    "y_pred = knnr.predict(X_test)\n",
    "stop=datetime.now()\n",
    "print(stop-start)\n",
    "\n",
    "adj_y_test = np.exp(y_test) - shift\n",
    "adj_y_pred = np.exp(y_pred) - shift\n",
    "\n",
    "print('The mae of the model prediction is: ', mean_absolute_error(adj_y_test, adj_y_pred))\n",
    "\n",
    "print('\\nBest params:\\n{}'.format(knnr.get_params()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# this code is based on the example in the sklearn documentation\n",
    "\n",
    "from sklearn.linear_model import ElasticNetCV\n",
    "\n",
    "enr = ElasticNetCV(cv=5, max_iter=4000, random_state=42)\n",
    "\n",
    "start=datetime.now()\n",
    "enr.fit(X_train, y_train)\n",
    "stop=datetime.now()\n",
    "print(stop-start)\n",
    "\n",
    "start=datetime.now()\n",
    "y_pred = enr.predict(X_test)\n",
    "stop=datetime.now()\n",
    "print(stop-start)\n",
    "\n",
    "adj_y_test = np.exp(y_test) - shift\n",
    "adj_y_pred = np.exp(y_pred) - shift\n",
    "\n",
    "print('The mae of the model prediction is: ', mean_absolute_error(adj_y_test, adj_y_pred))\n",
    "\n",
    "print('\\nParameters:\\n{}'.format(enr.get_params()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}