Linear_regression_model_using_tensor_flow_boston_housing.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Linear regression model using tensor flow_boston_housing",
      "provenance": [],
      "collapsed_sections": [],
      "authorship_tag": "ABX9TyMA7sY0ZFuad1IOF05rTFuA",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/SeungHuiHan/Example-of-linear-regression-using-tensor-flow/blob/master/Linear_regression_model_using_tensor_flow_boston_housing.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "KrwyrMhNKcJF"
      },
      "outputs": [],
      "source": [
        "#선형회귀 모델 구축\n",
        "\n",
        "import tensorflow as tf\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "import math\n",
        "from tensorflow.keras.datasets import boston_housing\n",
        " #한글 폰트 설정\n",
        "plt.rc('font', family='NanumBarunGothic')"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#시각화 라이브러리에 사용되는 한글 폰트 깨짐 방지\n",
        "!sudo apt-get install -y fonts-nanum\n",
        "!sudo fc-cache -fv\n",
        "!rm ~/.cache/matplotlib -rf"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "c_VbqFFoK1eu",
        "outputId": "6835c48b-c12b-4958-d718-db3ee9e8a184"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Reading package lists... Done\n",
            "Building dependency tree       \n",
            "Reading state information... Done\n",
            "The following packages were automatically installed and are no longer required:\n",
            "  cuda-command-line-tools-10-0 cuda-command-line-tools-10-1\n",
            "  cuda-command-line-tools-11-0 cuda-compiler-10-0 cuda-compiler-10-1\n",
            "  cuda-compiler-11-0 cuda-cuobjdump-10-0 cuda-cuobjdump-10-1\n",
            "  cuda-cuobjdump-11-0 cuda-cupti-10-0 cuda-cupti-10-1 cuda-cupti-11-0\n",
            "  cuda-cupti-dev-11-0 cuda-documentation-10-0 cuda-documentation-10-1\n",
            "  cuda-documentation-11-0 cuda-documentation-11-1 cuda-gdb-10-0 cuda-gdb-10-1\n",
            "  cuda-gdb-11-0 cuda-gpu-library-advisor-10-0 cuda-gpu-library-advisor-10-1\n",
            "  cuda-libraries-10-0 cuda-libraries-10-1 cuda-libraries-11-0\n",
            "  cuda-memcheck-10-0 cuda-memcheck-10-1 cuda-memcheck-11-0 cuda-nsight-10-0\n",
            "  cuda-nsight-10-1 cuda-nsight-11-0 cuda-nsight-11-1 cuda-nsight-compute-10-0\n",
            "  cuda-nsight-compute-10-1 cuda-nsight-compute-11-0 cuda-nsight-compute-11-1\n",
            "  cuda-nsight-systems-10-1 cuda-nsight-systems-11-0 cuda-nsight-systems-11-1\n",
            "  cuda-nvcc-10-0 cuda-nvcc-10-1 cuda-nvcc-11-0 cuda-nvdisasm-10-0\n",
            "  cuda-nvdisasm-10-1 cuda-nvdisasm-11-0 cuda-nvml-dev-10-0 cuda-nvml-dev-10-1\n",
            "  cuda-nvml-dev-11-0 cuda-nvprof-10-0 cuda-nvprof-10-1 cuda-nvprof-11-0\n",
            "  cuda-nvprune-10-0 cuda-nvprune-10-1 cuda-nvprune-11-0 cuda-nvtx-10-0\n",
            "  cuda-nvtx-10-1 cuda-nvtx-11-0 cuda-nvvp-10-0 cuda-nvvp-10-1 cuda-nvvp-11-0\n",
            "  cuda-nvvp-11-1 cuda-samples-10-0 cuda-samples-10-1 cuda-samples-11-0\n",
            "  cuda-samples-11-1 cuda-sanitizer-11-0 cuda-sanitizer-api-10-1\n",
            "  cuda-toolkit-10-0 cuda-toolkit-10-1 cuda-toolkit-11-0 cuda-toolkit-11-1\n",
            "  cuda-tools-10-0 cuda-tools-10-1 cuda-tools-11-0 cuda-tools-11-1\n",
            "  cuda-visual-tools-10-0 cuda-visual-tools-10-1 cuda-visual-tools-11-0\n",
            "  cuda-visual-tools-11-1 default-jre dkms freeglut3 freeglut3-dev\n",
            "  keyboard-configuration libargon2-0 libcap2 libcryptsetup12\n",
            "  libdevmapper1.02.1 libfontenc1 libidn11 libip4tc0 libjansson4\n",
            "  libnvidia-cfg1-510 libnvidia-common-460 libnvidia-common-510\n",
            "  libnvidia-extra-510 libnvidia-fbc1-510 libnvidia-gl-510 libpam-systemd\n",
            "  libpolkit-agent-1-0 libpolkit-backend-1-0 libpolkit-gobject-1-0 libxfont2\n",
            "  libxi-dev libxkbfile1 libxmu-dev libxmu-headers libxnvctrl0 libxtst6\n",
            "  nsight-compute-2020.2.1 nsight-compute-2022.1.0 nsight-systems-2020.3.2\n",
            "  nsight-systems-2020.3.4 nsight-systems-2021.5.2 nvidia-dkms-510\n",
            "  nvidia-kernel-common-510 nvidia-kernel-source-510 nvidia-modprobe\n",
            "  nvidia-settings openjdk-11-jre policykit-1 policykit-1-gnome python3-xkit\n",
            "  screen-resolution-extra systemd systemd-sysv udev x11-xkb-utils\n",
            "  xserver-common xserver-xorg-core-hwe-18.04 xserver-xorg-video-nvidia-510\n",
            "Use 'sudo apt autoremove' to remove them.\n",
            "The following NEW packages will be installed:\n",
            "  fonts-nanum\n",
            "0 upgraded, 1 newly installed, 0 to remove and 39 not upgraded.\n",
            "Need to get 9,604 kB of archives.\n",
            "After this operation, 29.5 MB of additional disk space will be used.\n",
            "Get:1 http://archive.ubuntu.com/ubuntu bionic/universe amd64 fonts-nanum all 20170925-1 [9,604 kB]\n",
            "Fetched 9,604 kB in 1s (8,339 kB/s)\n",
            "debconf: unable to initialize frontend: Dialog\n",
            "debconf: (No usable dialog-like program is installed, so the dialog based frontend cannot be used. at /usr/share/perl5/Debconf/FrontEnd/Dialog.pm line 76, <> line 1.)\n",
            "debconf: falling back to frontend: Readline\n",
            "debconf: unable to initialize frontend: Readline\n",
            "debconf: (This frontend requires a controlling tty.)\n",
            "debconf: falling back to frontend: Teletype\n",
            "dpkg-preconfigure: unable to re-open stdin: \n",
            "Selecting previously unselected package fonts-nanum.\n",
            "(Reading database ... 155113 files and directories currently installed.)\n",
            "Preparing to unpack .../fonts-nanum_20170925-1_all.deb ...\n",
            "Unpacking fonts-nanum (20170925-1) ...\n",
            "Setting up fonts-nanum (20170925-1) ...\n",
            "Processing triggers for fontconfig (2.12.6-0ubuntu2) ...\n",
            "/usr/share/fonts: caching, new cache contents: 0 fonts, 1 dirs\n",
            "/usr/share/fonts/truetype: caching, new cache contents: 0 fonts, 3 dirs\n",
            "/usr/share/fonts/truetype/humor-sans: caching, new cache contents: 1 fonts, 0 dirs\n",
            "/usr/share/fonts/truetype/liberation: caching, new cache contents: 16 fonts, 0 dirs\n",
            "/usr/share/fonts/truetype/nanum: caching, new cache contents: 10 fonts, 0 dirs\n",
            "/usr/local/share/fonts: caching, new cache contents: 0 fonts, 0 dirs\n",
            "/root/.local/share/fonts: skipping, no such directory\n",
            "/root/.fonts: skipping, no such directory\n",
            "/var/cache/fontconfig: cleaning cache directory\n",
            "/root/.cache/fontconfig: not cleaning non-existent cache directory\n",
            "/root/.fontconfig: not cleaning non-existent cache directory\n",
            "fc-cache: succeeded\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#데이터 불러오기\n",
        "(train_X, train_Y), (test_X, test_Y) = boston_housing.load_data() #X는 가격 예측을 위한 데이터(독립변수) Y는 주택 가격(종속변수)"
      ],
      "metadata": {
        "id": "cJ6szsQjMWXQ",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "10830b8b-f1d7-40d4-f4ea-667997b39f73"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/boston_housing.npz\n",
            "57344/57026 [==============================] - 0s 0us/step\n",
            "65536/57026 [==================================] - 0s 0us/step\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "train_X.shape #404개의 train 샘플 13가지 수치"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vqG-ZC7cpKjs",
        "outputId": "0c0ec4ea-871d-4077-835b-4386add354f9"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(404, 13)"
            ]
          },
          "metadata": {},
          "execution_count": 3
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "test_X.shape  #102개의 test 샘플 13가지 수치\n",
        "#train데이터와 test데이터 8:2<-가장 이상적"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "KspKfxTNpUBt",
        "outputId": "784e98e1-aab2-4b7c-9656-7b25551a0e4f"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(102, 13)"
            ]
          },
          "metadata": {},
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "**2022.01.03실습코드 Data Set Characteristics:**  \n",
        "\n",
        "    :Number of Instances: 506<-여기서 404개는 train_data, 102개는 test_data \n",
        "\n",
        "    :Number of Attributes: 13 numeric/categorical predictive. Median Value (attribute 14) is usually the target.\n",
        "\n",
        "    :Attribute Information (in order):\n",
        "        - CRIM     per capita crime rate by town\n",
        "        - ZN       proportion of residential land zoned for lots over 25,000 sq.ft.\n",
        "        - INDUS    proportion of non-retail business acres per town\n",
        "        - CHAS     Charles River dummy variable (= 1 if tract bounds river; 0 otherwise)\n",
        "        - NOX      nitric oxides concentration (parts per 10 million)\n",
        "        - RM       average number of rooms per dwelling\n",
        "        - AGE      proportion of owner-occupied units built prior to 1940\n",
        "        - DIS      weighted distances to five Boston employment centres\n",
        "        - RAD      index of accessibility to radial highways\n",
        "        - TAX      full-value property-tax rate per $10,000\n",
        "        - PTRATIO  pupil-teacher ratio by town\n",
        "        - B        1000(Bk - 0.63)^2 where Bk is the proportion of black people by town\n",
        "        - LSTAT    % lower status of the population\n",
        "        - MEDV     Median value of owner-occupied homes in $1000's\n",
        " "
      ],
      "metadata": {
        "id": "OmyL0xQ0hWtf"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#데이터 전처리(정규화)\n",
        "#데이터의 단위가 다른 경우는 비교하기 힘드므로(모델학습에 문제가 발생할 수 있음),\n",
        "# 일반적으로 전처리(Preprocessing) 과정에서 데이터를 정규화해줌 0~1범위\n",
        "x_mean = train_X.mean(axis=0) #평균\n",
        "x_std = train_X.std(axis=0)   #분산\n",
        "train_X -= x_mean #train_X=train_X - x_mean\n",
        "train_X /= x_std  #train_X=train_X / x_std\n",
        "test_X -= x_mean  #test_X=test_X - x_mean\n",
        "test_X /= x_std   #test_X=test_X / x_std\n",
        "\n",
        "y_mean = train_Y.mean(axis=0) #평균\n",
        "y_std = train_Y.std(axis=0)   #분산\n",
        "train_Y -= y_mean #train_Y=train_Y - y_mean\n",
        "train_Y /= y_std  #train_Y=train_Y / y_std\n",
        "test_Y -= y_mean  #test_Y=test_Y - y_mean\n",
        "test_Y /= y_std   #test_Y=test_Y / y_std"
      ],
      "metadata": {
        "id": "fek3LoeaMjl1"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(train_X[0])\n",
        "print(train_Y[0])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "UuZWq0l0wSTp",
        "outputId": "3cf464cc-1a13-43b2-eb3f-a14bcaf42b5a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[-0.27224633 -0.48361547 -0.43576161 -0.25683275 -0.1652266  -0.1764426\n",
            "  0.81306188  0.1166983  -0.62624905 -0.59517003  1.14850044  0.44807713\n",
            "  0.8252202 ]\n",
            "-0.7821526033779157\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#Boston Housing Dataset 회귀 모델 생성\n",
        "model = tf.keras.Sequential([\n",
        "    tf.keras.layers.Dense(units=64, activation='relu', input_shape=(train_X.shape[1],)), #샘플 개수가 적기 때문에 64개의 유닛을 가진 2개의 은닉 층으로 작은 네트워크를 구성\n",
        "    tf.keras.layers.Dense(units=64, activation='relu'),                         #일반적으로 훈련 데이터의 개수가 적을수록 과대적합이 더 쉽게 일어나므로 작은 모델을 사용하는 것이 과대적합을 피하는 한 방법\n",
        "    \n",
        "    tf.keras.layers.Dense(units=1) #하나의 연속적인 값을 예측하는 회귀를 위한 선형층 네트워크가 어떤 범위의 값이라도 예측하도록 자유롭게 학습되도록 활성화 함수 적용X \n",
        "])                                 #마지막 레이어는 주택가격인 Y값 1개만 예측하면 되기 때문에 뉴런의 수가 1개\n",
        "\n",
        "model.compile(optimizer=tf.keras.optimizers.Adam(lr=0.07), loss='mse') #Adam옵티마이저:학습률을 줄여나가고 속도를 계산하여 학습의 갱신도강도를 적응적으로 조정해나가는 방법\n",
        "\n",
        "model.summary()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "qu-Q_aTjwU0E",
        "outputId": "d4df1d1c-7619-4f30-c6df-7ae7b6db7036"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " dense (Dense)               (None, 64)                896       \n",
            "                                                                 \n",
            " dense_1 (Dense)             (None, 64)                4160      \n",
            "                                                                 \n",
            " dense_2 (Dense)             (None, 1)                 65        \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 5,121\n",
            "Trainable params: 5,121\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/keras/optimizer_v2/adam.py:105: UserWarning: The `lr` argument is deprecated, use `learning_rate` instead.\n",
            "  super(Adam, self).__init__(name, **kwargs)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "history = model.fit(train_X, train_Y, epochs=40, batch_size=32, \n",
        "                    callbacks=[tf.keras.callbacks.EarlyStopping(patience=5, monitor='val_loss')],validation_split=0.25,verbose=1) \n",
        "#EarlyStopping:모델을 더 이상 학습을 못할 경우(loss등의 개선이 없을 경우), 학습 도중 미리 학습을 종료시키는 콜백함수\n",
        "#val_loss가 더이상 감소되지 않을 경우 EarlyStopping을 적용\n",
        "#GPU/RAM에서 일반적으로 작동하는 방식이 2의 제곱이기 때문에 2의 제곱이 아닌 다른 숫자를 설정하면 비효율적이며 성능 저하가 일어날 수 있음"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xpdgWCZP5sxv",
        "outputId": "a46a1fbd-55fe-48a5-f550-325662ab8db2"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/40\n",
            "10/10 [==============================] - 2s 18ms/step - loss: 3.7667 - val_loss: 1.3254\n",
            "Epoch 2/40\n",
            "10/10 [==============================] - 0s 7ms/step - loss: 0.5844 - val_loss: 0.5514\n",
            "Epoch 3/40\n",
            "10/10 [==============================] - 0s 5ms/step - loss: 0.3513 - val_loss: 0.4428\n",
            "Epoch 4/40\n",
            "10/10 [==============================] - 0s 5ms/step - loss: 0.2905 - val_loss: 0.3314\n",
            "Epoch 5/40\n",
            "10/10 [==============================] - 0s 6ms/step - loss: 0.2204 - val_loss: 0.3072\n",
            "Epoch 6/40\n",
            "10/10 [==============================] - 0s 5ms/step - loss: 0.1851 - val_loss: 0.2887\n",
            "Epoch 7/40\n",
            "10/10 [==============================] - 0s 7ms/step - loss: 0.1715 - val_loss: 0.3222\n",
            "Epoch 8/40\n",
            "10/10 [==============================] - 0s 5ms/step - loss: 0.2068 - val_loss: 0.2055\n",
            "Epoch 9/40\n",
            "10/10 [==============================] - 0s 6ms/step - loss: 0.1221 - val_loss: 0.1776\n",
            "Epoch 10/40\n",
            "10/10 [==============================] - 0s 6ms/step - loss: 0.1054 - val_loss: 0.1894\n",
            "Epoch 11/40\n",
            "10/10 [==============================] - 0s 6ms/step - loss: 0.1024 - val_loss: 0.1695\n",
            "Epoch 12/40\n",
            "10/10 [==============================] - 0s 5ms/step - loss: 0.1067 - val_loss: 0.2167\n",
            "Epoch 13/40\n",
            "10/10 [==============================] - 0s 5ms/step - loss: 0.1225 - val_loss: 0.1778\n",
            "Epoch 14/40\n",
            "10/10 [==============================] - 0s 5ms/step - loss: 0.1018 - val_loss: 0.2658\n",
            "Epoch 15/40\n",
            "10/10 [==============================] - 0s 5ms/step - loss: 0.1044 - val_loss: 0.1395\n",
            "Epoch 16/40\n",
            "10/10 [==============================] - 0s 5ms/step - loss: 0.1006 - val_loss: 0.2390\n",
            "Epoch 17/40\n",
            "10/10 [==============================] - 0s 6ms/step - loss: 0.1141 - val_loss: 0.1571\n",
            "Epoch 18/40\n",
            "10/10 [==============================] - 0s 6ms/step - loss: 0.1002 - val_loss: 0.2236\n",
            "Epoch 19/40\n",
            "10/10 [==============================] - 0s 6ms/step - loss: 0.1318 - val_loss: 0.1965\n",
            "Epoch 20/40\n",
            "10/10 [==============================] - 0s 5ms/step - loss: 0.1061 - val_loss: 0.1628\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#회귀 모델 학습 결과 시각화\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "plt.plot(history.history['loss'], 'b-', label='loss')\n",
        "plt.plot(history.history['val_loss'], 'r--', label='val_loss')\n",
        "plt.xlabel('Epoch')\n",
        "plt.legend()\n",
        "plt.show()\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 279
        },
        "id": "yrgzQiW88eEf",
        "outputId": "0ec3e0e7-0af2-4146-b1c4-f7b673fc2900"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXMAAAEGCAYAAACXVXXgAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3de3xU9Z3/8dcnFwhJAJEEgkIEbRGwWq1B1IIiWa3alq1a76tFtNSy3n5ubWldV/2ttdrWWuuuWuqVqrVeqlUutnhBoKASC4oKtFULBAET5BqIgeS7f3zPkMlkIEMyk8mceT8fj/OYmXPOzHzmZPKe73zP+Z4x5xwiIpLZctJdgIiIdJzCXEQkBBTmIiIhoDAXEQkBhbmISAjkpeNJS0pK3ODBg9Px1CIiGeutt96qdc6VxluWljAfPHgwVVVV6XhqEZGMZWYr97RM3SwiIiGgMBcRCQGFuYhICKSlz1xEZOfOnVRXV1NfX5/uUrqcgoICBg4cSH5+fsL3UZiLSFpUV1fTs2dPBg8ejJmlu5wuwznHhg0bqK6uZsiQIQnfT90sIpIW9fX19O3bV0Eew8zo27fvPn9jUZiLSNooyONrz3bJqDCfPx9++EPQWXtFRFrKqDCvqoLbboNNm9JdiYiEwZw5czjvvPPSXUZSZFSYl5X5y3Xr0luHiEhXk1FHs0SH+fDh6a1FRJLnmmtgyZLkPuaRR8Ivf5nYuvPnz+eGG27AzOjevTt33nknQ4cO5eyzz2bt2rXk5eUxZcoUDj/8cM477zxycnLo0aMH9913HwcffHByC2+njArz/v39pVrmIpIszjkuueQS5s6dy4ABA3jnnXe47LLLmDlzJkuXLmXJkiUUFBRQXV3N3/72N8rKynj66afZtm0bDQ0N6S5/t4wKc3WziIRToi3oVKipqaGoqIgBAwYAcMQRR/Dhhx/Sq1cvnnvuOW6++WbMjGuuuYbKykp27drFNddcw8CBA7n66qvTV3iMjOoz328/6NZNYS4iyVNSUkJdXR1r164F4J133mHIkCHU1tYyYMAAbr/9diorK7nqqqv45z//ybhx47jrrrvYtGkT999/f5qrb5ZRLXMz3zpXmItIsuTk5PDQQw9x/vnnk5OTQ15eHvfffz+bN2/m8ssvp6Ghgc8++4wbbriB999/nwsvvJBu3brR1NTEQw89lO7yd8uoMAeFuYgkz9ixYxk7dizgD1OMNXv27FbzTj/99BRX1T4Z1c0CPszXr093FSIiXUtGhrla5iIiLWVkmNfUQGNjuisREek6MjLMm5p8oIuIiNdmmJvZfmb2pJktNLPXzezamOWDzWydmc0JpumpK1fHmouIxJPI0SzdgZucc++bWR6wzMymOedqo9Z50Tk3ISUVxlCYi4i01mbL3Dm33jn3fnCzFNgF1MWsVmlm883sFTMbn+wio2lIv4hIawkfZ25mtwGTgB8453ZELVoJlDvnnJmVA7PNbIVzbkXM/ScF96e8vLzdBSvMRURaS3gHqHNuCjAIuNjMjoma75zzPxfhnFsFvAQcFuf+U51zFc65itLS0nYXXFQEPXsqzEWkczz88MNMmTJln5d1tjZb5mZ2KPCpc64G2A5sBvqYWS/n3BYzGwqsds7tMLM+wBjgtlQWrWPNRUIoGInZwjnnwOTJsH07xBt5OWGCn2pr4ZvfbLkszojOMEukZb4L+LWZvQwsAJYBJcCjwfID8F0rrwEzgOudc6tTUWyERoGKSEedc845vPHGGwDs3LmTESNGcPnllzNu3DiOPvpobrzxxn16vOeee44TTjiByspKzjjjDNauXcuWLVs4+eSTGT16NCeddBKLFi1i8eLFjBw5khNOOIGvf/3rbNy4MSmvp82WuXPuA+DMOIseC5bPAUYnpZoElZXB0qWd+YwiknJ7a0kXFu59eUnJPrfEr7zySqZOncqoUaN49tlnGT9+PBdddBHDhg1j5cqVjBw5kptvvjmhx/r000+ZMmUKVVVVFBcX88ILL/C9732PG2+8kW3btjF37lycc9TW1jJr1izGjBnDL37xC2pqaujWrds+1b0nGTdoCNTNIiIdN2bMGJYvX86WLVt44IEHmDx5Mrfddhs/+MEPmDlzJjk5icfjP/7xD4YNG0ZxcTEAo0ePZsmSJQwdOpQ777yTKVOmcOutt1JcXMzEiRMZNWoUV155JX/4wx8oLCxMyuvJ2DDftAnq69NdiYhksokTJ3LTTTdRXFzMM888w/Dhw/n5z3/OySefzObNmxN+nEMOOYTly5ezbds2AObNm8eRRx7JmjVrOOKII7jjjjs48MADueWWW/jggw8455xzuPvuu5k7dy6zZs1KymvJuFPgQvPAofXr4aCD0luLiGSuCy64gEGDBvHkk0/Sr18/zj//fGbMmMFRRx3Ffvvtl/Dj9O3bl5/85CecdtppdOvWjeLiYu69916qq6u54IILMDMaGhq46667mDdvHhdffDHdunWjoKCAY489NimvxYKjCjtVRUWFq6qqavf9Z86Er34VXn8dRo1KYmEi0mmWLVvGcP0y+x7F2z5m9pZzriLe+hnZMtfAIRHpbFOnTuXxxx/fffvoo4/mjjvuSGNFLWVkmOv8LCLS2SZNmsSkSZPSXcYeZeQO0H79/KXCXCSzpaObNxO0Z7tkZJjn5/vDShXmIpmroKCADRs2KNBjOOfYsGEDBQUF+3S/jOxmAY0CFcl0AwcOpLq6mhr90kwrBQUFDBw4cJ/uk9Fhrpa5SObKz89nyJAh6S4jNDKymwUU5iIi0TI+zNXdJiKS4WG+Ywds3ZruSkRE0i+jwxzU1SIiAhkc5hoFKiLSLGPDXC1zEZFmCnMRkRDI2DDff3/Iy1OYi4hABod5To7vN9coUBGRBMLczPYzsyfNbKGZvW5m18ZZ58dmtiBYZ2xKKo1DA4dERLxEhvN3B25yzr1vZnnAMjOb5pyrBTCzccCRzrnjzewA4BUz+4JzblcK6wZ8mK9dm+pnERHp+tpsmTvn1jvn3g9ulgK7gLqoVSqBp4J1PwZWAocmuc641DIXEfES7jM3s9uA94BfOOd2RC0qBWqjbtcG82LvP8nMqsysKllnSYucObGpKSkPJyKSsRIOc+fcFGAQcLGZHRO1qA7oHXW7N7Axzv2nOucqnHMVpaWtsr5dysqgsRE2bEjKw4mIZKxEdoAeamaR9N0ObAb6mFmvYN7LwPhg3RJ8F8uKFNTaio41FxHxEmmZ7wJ+bWYvAwuAZUAJ8GiwfAaw3swWANOBq51z9akoNpaG9IuIeG0ezeKc+wA4M86ix4LlDrgqyXUlRC1zEREvYwcNgcJcRCQio8O8uBgKCxXmIiIZHeZm+mFnERHI8DAHDRwSEQGFuYhIKCjMRURCIBRhvmEDNDSkuxIRkfQJRZgDfPJJeusQEUmnjA9zjQIVEQlBmGvgkIiIwlxEJBQyPszVzSIiEoIw794d+vTRKFARyW4ZH+agY81FRBTmIiIhoDAXEQkBhbmISAiEJsy3bfOTiEg2Ck2Yg45oEZHs1WaYm1mRmd1jZm+a2SIzuzVm+WAzW2dmc4JpeurKjU/HmotItmvzB52B/YDHnXOTzSwHWGZmv3LORUfni865CSmpMAEaBSoi2a7Nlrlzbo1zbn5wswhoADbFrFZpZvPN7BUzG5/sItuiMBeRbJdIyxwAM8sFpgHXOefqoxatBMqdc87MyoHZZrbCObci5v6TgEkA5eXlHa88SkkJ5OSoz1xEsldCO0DNLB94FHjCOfdi9DIXCK6vAl4CDot9DOfcVOdchXOuorS0tOOVR8nNhX791DIXkeyVyA7QbsATwPPOud8H83LNrFdwfaiZ9Qiu9wHGAItSV3J8OtZcRLJZIt0slwFjgb5m9p1g3mxgFDAeOAB40MwagXzgeufc6hTUulcKcxHJZm2GuXPuHuCevSyfA4xOYk3tUlYG772X7ipERNIjFIOGoLll7nvvRUSyS6jCfOdO2Lgx3ZWIiHS+0IS5RoGKSDYLTZhr4JCIZDOFuYhICIQuzDUKVESyUWjCvHdv/+POapmLSDYKTZibaeCQiGSv0IQ5KMxFJHspzEVEQkBhLiISAqEL85oa2LUr3ZWIiHSuUIV5//7+3Cw1NemuRESkc4UqzDVwSESylcJcRCQEQhnmGgUqItkmVGGuMyeKSLYKVZgXFkKvXgpzEck+oQpz0LHmIpKd2gxzMysys3vM7E0zW2Rmt8ZZ58dmtsDMFprZ2JRUmiCFuYhko0Ra5vsBjzvnjgFGAWeZWVlkoZmNA450zh0PnAXcZ2Zt/lB0qijMRSQbtRnmzrk1zrn5wc0ioAHYFLVKJfBUsO7HwErg0NjHMbNJZlZlZlU1KRzVozAXkWyUcJ+5meUC04DrnHP1UYtKgdqo27XBvBacc1OdcxXOuYrS0laLk6asDDZvhh07UvYUIiJdTkJhbmb5wKPAE865F2MW1wG9o273BjYmp7x9Fzk8Uceai0g2SWQHaDfgCeB559zvg3m5ZtYrWOVlYHwwvwTfxbIiNeW2TaNARSQbJbKj8jJgLNDXzL4TzJuN3xk6HpgBnGJmC/AfDlfHdMN0Ko0CFZFs1GaYO+fuAe7Zy3IHXJXMojpCLXMRyUahGzRUWup/D1RhLiLZJHRhnp8PJSUKcxHJLqELc9Cx5iKSfRTmIiIhoDAXEQmBUIe5c+muRESkc4QyzPv3h/p62LIl3ZWIiHSOUIa5jjUXkWyjMBcRCYFQh7mG9ItItgh1mKtlLiLZIpRh3qePHwmqMBeRbBHKMM/J8Ue0KMxFJFuEMsxBA4dEJLsozEVEQkBhLiISAqEO808+gcbGdFciIpJ6oQ3z/v19kG/YkO5KRERSL7RhrmPNRSSbtBnmZnaomS0wsyfiLBtsZuvMbE4wTU9NmftOo0BFJJu0+YPOwCjgV8A39rD8RefchKRVlCRqmYtINmmzZe6cmwbsLRIrzWy+mb1iZuP3tJKZTTKzKjOrqqmpaU+t+0RhLiLZJJGW+d6sBMqdc87MyoHZZrbCObcidkXn3FRgKkBFRUXKfzaiuBiKihTmIpIdOrQD1AWC66uAl4DDklFYMuhYcxHJFvsc5maWa2a9gutDzaxHcL0PMAZYlNwS209hLiLZoj0t8/OAR4PrB+C7Vl4DZgDXO+dWJ6u4jlKYi0i2SKjP3Dk3B5gTXH8MeCxq/ujUlNZxZWXw6qvprkJEJPVCO2gI/CjQTz+Fzz5LdyUiIqkV6jCPHJ74ySfprUNEJNWyIsw1ClREwi6zwtw5eP11qK9PaHUNHBKRbJFZYT5vHhx3HPzxjwmtrjAXkWyRWWE+ejSUl8ODDya0er9+/lJhLiJhl1lhnpMDl1wCs2fDypVtrt69O+y/v8JcRMIvs8IcYMIEf/nIIwmtroFDIpINMi/MBw+Gykp47rmEVleYi0g26OhZE9PjgQeaO8TbUFYGb7yR4npERNIsM8O8vDzhVfv3V8tcRMIv87pZIv78ZzjiCNi4ca+rlZVBXR1s29ZJdYmIpEHmhnlJCSxdCr/73V5X0yhQEckGmRvmRx0FX/xim8eca+CQiGSDzA1zM5g4Ed56C95+e4+rKcxFJBtkbpgDXHghdOsGDz20x1UU5iKSDTLzaJaIvn3hv/8bvvCFva6Sm6swF5Fwy+wwB/j+9/e6ODfXH5KuMBeRMMvsbpaIVavg0Uf3uFijQEUk7BIKczM71MwWmNkTe1j+42D5QjMbm9QKE3H//XDxxbA6/m9JK8xFJOwSbZmPAn4Vb4GZjQOOdM4dD5wF3Gdmndt9M2GC/+GKadPiLlaYi0jYJRTmzrlpwJ7isBJ4KljvY2AlcGjsSmY2ycyqzKyqpqamneXuwcEHw7hx/pjzpqZWi/v394OGnEvu04qIdBXJ6DMvBWqjbtcG81pwzk11zlU45ypKS1st7riJE+HDD2Hu3FaLyspg5842R/6LiGSsZIR5HdA76nZvoPNj88wz/RD/OAOIdKy5iIRdu/q2zSwXKHLObQFeBi4CHjOzEnwXy4rklZigHj38rw8VFrZaFB3mI0Z0cl0iIp2gvS3z84DIsYAzgPVmtgCYDlztnKtPRnH7LBLk9S2fXi1zEQm7hFvmzrk5wJzg+mPAY8F1B1yVgtra5/LLYfHiFr9IoTAXkbALx6ChaMOGwZtv+tPjBnr1goIChbmIhFf4wvzCCyE/v8XJt8x0rLmIhFv4wry0FMaPh9/+Fhoads9WmItImIUvzAEuvRRqa+GFF3bPUpiLSJiFM8xPOQXuuQdOPHH3LP2ws4iEWeafAjee3Fz47ndbzCor8431XbsgL5yvWkSyWDhb5uBPxPLQQ/D444APc+cg2aeFERHpCsIb5mbwyCNw443gnI41F5FQC2+Ygz/51j/+AfPmKcxFJNTCHeZnnQU9e8KDDyrMRSTUwh3mRUVw/vnw1FP077EFUJiLSDiFO8zBd7UMG0aPDdX07q0wF5FwCv9BeqNGwVtvARo4JCLhFf6WecTmzRzat1ZhLiKhlB1hvn07lJdz6YbbFeYiEkrZEeaFhVBZyUmrplG7dme6qxERSbrsCHOASy+l545PGLN1BvPnp7sYEZHkyp4w/8pX2NVvANcUTuWkk+Duu/3wfhGRMMieMM/LI++ySzhx50uMP6Weq66Ciy9sZPv2dBcmItJxCYW5mV1hZgvN7HUzOzdm2WAzW2dmc4JpempKTYIbbsD++EeeeqGA//7/jv/3u5G8PuAM1t7zLHz2WbqrExFptzbD3MwOASYCJwInAzeZWZ+Y1V50zo0Npq+loM7kKCiA004jJwf+8/sN9DnzJA7bupAB/34mDSUDYPJkePfddFcpIrLPEmmZjwOed841OOe2AnOB42PWqTSz+Wb2ipmNj/cgZjbJzKrMrKqmK5yHtnt3hjxzB9tXVHPFIbN4etup7PzNQzQtftsv//RT+Oij9NYoIpKgRMK8FKiNul0bzItYCZQ750YDE4CfmdmhsQ/inJvqnKtwzlWUlpbGLk6bIZ/P42dLT2X2hMfpu2s93/jtWXz6KfDAA3DwwXDCCXD//bB5c7pLFRHZo0TCvA7oHXW7N7AxcsMFguurgJeAw5JZZKr16AEPPgg/vbcXL84poKIC3j3sXPjxj+GTT+Db3/a/O3fBBdDUlO5yRURaSSTMXwZON7NcM+sBjAWqzKwXgJkNDeYT9KWPARalqN6UMYPLL4d586ChAUaeVc60gT+CZcvgzTd9oJtBTrDJpk5V/7qIdBnmEjjY2sx+CHwDcMBU4DPgXOfceDMbC9wCNAL5wE+ccy/s7fEqKipcVVVVB0tPnU8+gXPPhTlz/D7RO++Ebt2iVti4EQYM8EfAHH00TJjgT7Xbt2+aKhaRbGBmbznnKuIuSyTMk62rhzn4H37+0Y/gZz+DY4+Fp5+GAw+MWqGmxv++6MMPw5IlkJ8Pv/89nHFGukoWkZDbW5hnz6ChfZSXBz/9KTz1lO9N+dKXfEt9t9JSuPpqWLzYT5Mn+9PtAjz/PFx3nbphRKTTKMzb8M1v+i7zPn3gX/4FbrkFVqyI2Q965JHwy1/CAQf422+/7W8ffjiMHAn/+7/+UEcRkRRRN0uCtmzxP1r0zDP+ds+evrt85EioqPCXgwf7faRA626YL31p949k8NZbsP/+vnVfVBR1JxGRPVOfeZI4B++9B4sWQVWVv3z7bX/0C/j9nxUVzdPIkb6xbm8v8TtNTzoJdu5suTe1oKC5y+Y//gN27PCd9aWlfiop8ZdDh0K/fp3zQrdvh+pqWL3aT9XVvrYePeCVV+Cdd+Cii7TDV6STKcxTqKEBli714R4J+HffhcZGv7ysrDnYKypg5JE7Kf3rn3zLvbbWX9bUwOmnw9lnw5o1MHw4bN3a8ol+/nMfqH//u++b32+/ltMVV8C4cf538Z58svXyIUP814mdO5uDetWq5sD+4Q9h0CC4917f/x/rb3+Dz38err3WH97TvTuceaY/ZPPEE5sP2RSRlFGYd7IdO3zPSiTcq6pg+fLmU+4OHQpjxsDo0X465JA4PS319S3D/nOf8yNS16yB226DTZtaTrfcAv/6r/DaazB2bOuinn0WvvEN30/0zW+2XNanD8yYAccd53fmzpoFAwf6cB80yF8vKGhe/513/KjY3/7WP/cpp8Cf/pTMTSgicSjMu4CtW+Gvf4U33oD58/20MRhHW1bmQz0S8Ecc4Y+maZfGxtZBv3mzP77ygAN8q3z27JZhXVTUvufasQP+8AfIzYXzzvMfQJde6kfKnnqqny/p55z/kJ4+3e/N/8pX4Fvfgl690l2Z7COFeRfU1OQHl86b54N93jzf6wG+N+S445oD/phj/C/fdXlvvw0nn+y/SQwc6PcYT5wIBx2U7sqyT12d3zeTnw+33grXX++//pWXw8qVftDbypV+eWfWtH273wck7aIwzxCrVsFf/tIc8O++6xtV+fn+yJnRo33IH3yw/5/s06cLHgjT0AAvvOC7YSJdL0uXwmEZdbqe1Ghq8sOLS0o68NVrL1au9K3v6dPh1Vfhuef8N6Tly/1XwtNO8zvR33zTtyS+9S3/BrviCr/e6acn/9vU9u0wc6bfjzN9uv/mdvfd/n1yww3+eN8xY1p246XCtm3+srgYNmzw3ZH9+/uvxf37+/kZQGGeoTZuhAULmsN90aLmI2fAv//Ky33DN97lAQekJjMStfODVWyZ9izvjbuKlauMIQ/8J27rVnodPZShJ5TRY0iZ7+YpL09fkalQX+93VC9b5vdjdO/uhxL/13/5Zbm5/jUffLAP3OJif5hUXZ3fUV1Ssm+f0mvW+DCODFL7/Ofha1/zO6eHD9/7fT/+2O+d//hjX8/kyf7bVJ/Ynyxoh+9+F6ZN84Her5/fV/Nv/9a8b+bYY/0buqDA70T/ylf8eTQi4zXayzn44AP/z7NwoZ+WLoXf/Ma/tldegcrKlvcpLPT7k0491e/w+vWvW4Z9//6+/zPNoa8wD4n6ev+eXLnSt+Ijl5HrGza0XD8315+CIDrgBw70XaVFRf59WVzc+nphYWJZsmVLyzpiLz/+uOXgqge5hAt5jG7s3D1v9ZAT2Pz8axx2GNhXT/c7F8rKmqcvfcm3KAHWr4fevX0AbN/up8JCv15jI7z4YvP8yFRR4YNiyxZ/JM5nn/kN0KePP9b/xBP9156GBt+C7dPHT4kc/79pkw+iggKYO9cH9rJl/jz4kRf+zjt+8NiLL/p9FQcd5F/HRx/5EJ4zxz/PhAnwyCP+PsXFPtSHD/eniAD/h8/N9TXPmeNbuYMH+x3fTU0+KMeMga9+1e9h3xc7d/od5Hff7VsNhYW+1uNjf7ZgL3bs8DvOX30VfvUr/5p+8AP/9zz7bH8q6dhWf12dbyH/6U9+WrHCfzU9/ni/3ZYt8y33tg6B3b7dt3QKCvyRXjU1zYfx9uzpPzSOO85vo8MP98/797/7v8P69f4IsPXr4Tvf8dvuj3+Eyy7z/1DR+fjmm/6D74kn/BlVI/ucItMZZ/j3lnMp+8qsMM8SdXUtwz32enW1P+dMW8x8lsUL/Lw8/zirVvksi5af39zQjvdtYdAgyKWRN2fWsuiFdSyfs44VK7vzGmMZNAgeLZrE8Ly/s3/DOnI/Weef4Pzz/eAr8P8osYdsTp7sR9jGHr8fcd11/rwMW7bAiBG+yC1b/GM3NcHtt8P3v+9bcp/7XMsX06ePD+iLL/Yb8frrfWB88IEPmvXrfRfCaaf58Lv2Wh/Aw4fDsGH+csSI+HXF+uAD3zr/8EMf9B995EPhheCcdZWVvkUZUVLiuyxuu63tx94Xixf7c/n/7Gd+XMEzz/g3xPjxrb/m7djhP6SefNLXWVfn+8MXL445kVGCVq7098vL8x8EP/2pf+6KCt9qP+UU+PKX/WGwTz/tPwgWLvT7anbt8kdrPfusf6zHHvMt6REj2t91tGuX/2CIhP2Xv+w/HGbNgvvuaz6stzb4uYePP/b7Im65Be66q3XYX3ON/5bWAQpzAXzjtabGdx9Gprq6xK/X1fkGbGxrP3JZVrbvh5uvXu3zYNYseOkln9V5eX7/wNf+pZ5Tx9Yz4vj9MJw/7fCGDT5QCwv9NHy4by2BbzkVFjZ/vSgs9IEUr6+pqck/WW6u/5TautW3DjdubDldcIFvVS5ZAmed5VuBQ4Y0h/WZZ/pjS1Nt8eLmD5DjjvOvuTOOFho3zre2Bw3y3SYXXeQ/VHv1aj7MtW9fv23OOcd/00lG315jo29t//nP/u/yxhv+w3X9ev8mO/ZY3610zDG+JX/ccX5eOgay7djhWziHHOJrmznTn58pEvarV/t/nvr6Do/HUJhLRmho8N2cs2b5aelSP3/gQN+Vedpp/lt3th5R55z/ApKf34k7vhsbfav77rubvxnccov/lrJ9u+8WOemk1O+c2bTJD1w75hh/e+1a/y0ghc+7a5f/7Kiu9j1ia9b428XFvhenXz/flR653qPHXh6srq79hwBHUZhLRqqubm61z57d3MPSo4cP9N69mxuJkettXfbs6f//c3J8wzZyGX09+nJvDSnnfAN/167mqbFx79d37vSvY+tW39sTudzT9dh5jY3+m/r++zd3+0emtm737p1YwzDyuiJTY2OwC+D998l94TmaThgLxx9PXl7ztmvPh4tzvlEb++0v3jfCyOX27f6LWaT7r62pqMhvr9j66uqaAzo6rKNvr1vX+ofFcnL2/GNjkZCPDvh4oV9e7t+H7aEwl4y3c6dvtf/lL/4ElJGA27y55WVkSubbOjbgI8EcOWVDMpj5MIh8OPXs2fp6z56+52jrVr8NItPGjc3XI0fg7ek5evXyl/HCOjK1Z9uZ0SLcc3Pj387J8YFcV+enfXmuggL/Qf7ZZ/4xEhXpSSsu9vevrW29vwf8h92BB/pp4MDm69G3S0r8B1BNjT/KNHpav771vJqa1u+T//kf+Pd/T7z+aHsL8zQeuCaSuPx83x174oltr9vU5IMiOuAj1yOt28gUHWbxLmPnNTW1DKq8vL1fj3e7Z8/WYV1UlJzT2zQ0NHf3Rwd+JPQjIRb51hH7LaSt25EPgsi2iXyoJXq7satztCAAAAcPSURBVLF5t0a8I6niXRYX+/tE96g0NvpAj97/k8i0fbvvVo8X2IkedRg5OGDw4LbXbWry2z066I86ap//rAlRmEvo5OQ0B2a26dat+bDoMMvNzYy/cU6O//Do29cfWJPS50rtw4uISGdIKMzN7AozW2hmr5vZuXGW/9jMFgTrjE16lSIisldtdrOY2SHAROBYoDvwppn92Tm3MVg+DjjSOXe8mR0AvGJmX3DOJTA8RUREkiGRlvk44HnnXINzbiswF4ge51sJPAXgnPsYWAkcGvsgZjbJzKrMrKqmpqbjlYuIyG6JhHkpUBt1uzaYl+hyAJxzU51zFc65ilKdAlNEJKkSCfM6oHfU7d7Axn1YLiIiKZZImL8MnG5muWbWAxgLVJlZr6jl4wHMrATfxbIiBbWKiMgetLkD1Dn3rplNBxYADvgFPtDPxYf4DOAUM1uA/3C42jlXn7KKRUSklbQM5zezGvyO0vYooWUffVej+jpG9XVcV69R9bXfQc65uDsd0xLmHWFmVXs6N0FXoPo6RvV1XFevUfWlhkaAioiEgMJcRCQEMjHMp6a7gDaovo5RfR3X1WtUfSmQcX3mIiLSWia2zEVEJIbCXEQkBLpsmHfl0+6aWZGZ3WNmb5rZIjO7NWb5YDNbZ2Zzgml6Z9YXVcemqBrmmFlxzPK9buMU1nV1TF0fm9mJUcvTsv3M7NDgPfVE1Ly9vs/MLN/MpprZPDOba2Zf6Kz6zKzUzB4zszeCk9hdEec+Y83sn1Hb8v5OrC+hv2Nn/i/HqfEnMe/FT83soJj7dNo27BDnXJebgEOAvwLdgJ7AMqBP1PJxwIzg+gHAciCvE+s7EBgdXM/Bn76gLGr5YODhLrAd57R3G3dijd2AKqBHurcfcDFwHvBEou8z/Omh7wmuHwks6MT6DgO+EFzvAawj2A8WdZ+xwE1p2n5t/h07+385tsaYZSX4ke5p24Ydmbpqyzwpp91NFefcGufc/OBmEdAAxP5EbKWZzTezV8xsfGfVFuOwoLU418wujVnW1jbuLBcDzzjndsTM7/Tt55ybhg/E3TXQ9vusEngyWGcJ0NfMijqjPufce865d4ObfYFqF6RPjAuDbTnLzEanorZ49QXa+jt26v/yHmqMuAa4O53bsCO66m+AJnLa3YV7Wd4pzCwXmAZc51qej2YlUO6cc2ZWDsw2sxXOuc4+AVl/51yTmfUFZpjZP51zLwfLEjp1cSoF2+9y/AdLtK6y/RJ5n+1pO9altrRmwYfHNOCyOIvnOuc+H6x3ODDdzL7onIvz+/RJl8jfsav8L/cGTgVujLM4ndswYV21Zd7lT7trZvnAo/ivay9GL3OB4Poq4CX8V+JO5ZxrCi43AM8AX4xanPZtiP+6+yfn3JbomV1l+5HYNkrrdjSznsDTwM3BN4MWIu+B4PpS4F3g4M6oLcG/Y1d4HwJcAfzaOdcYuyCd23BfdNUw79Kn3TWzbsAT+G6K3wfzciP1mdnQoG7MrA8wBljUWfUFz3uQme0XXO8BfB2YF5lH/G38RifWZ8DVwC+D211q+wXivs/MrJuZFcZZ51Bgl3Nuc2cUF7QmnwNud869FjV/d31mdpiZ5QXXD8KH0PJOqi/u3zH6b00XOIV28M3mbOCRqHldYhvuiy7ZzeK6/ml3Lwvq6Wtm3wnmzQZGBfUdADxoZo1APnC9c251J9YH0At4OOjKyAfuBwqAOfjfbG21jZ1ze+pLTIUzgIXOuchvCJ5H89+3K2w/2MP7zMx+BAwArgQeAB4ws3mAAd/uxPquB4YBN/nPRgAuBL4VVd+IoL7PguUTnXPbO6m+uH9HM7uQrvO/DPAd4BHnXEPUvO/RNbZhwjQCVEQkBLpqN4uIiOwDhbmISAgozEVEQkBhLiISAgpzEZEQUJiLiISAwlxCKzhr35aYs+Jdm4THfdjMTk1GjSLJ0iUHDYkk0fvOubHpLkIk1dQyl6wSnJv6FTN72sz+YmYvmFkvM+tt/tzgL5s/N/llwfolZvZEMG9OVIt8TLDuUjO7II0vSQTQCFAJMTMbDLyDP297xKvABOCLzrktZvZT/Jn6IqeQvTs4idprwCXAD4HXnXP3BY/ZF7gDaHLOTTSzQcBM59zhnfOqROJTN4uEXYtuFvO/ZHNI1JkaFwIn439I4XcAzrmdZrYIOAKoAP4jcn/n3IbgPCh/CG6vDk4QJZJW6maRbDQschY8/LnUlwbTCbD79MYjo+bv/lEFM+sXXN0V9XiGSJqpZS5hN8LM5kTdfh4fvs+a2f7AR8BDQHfgXjP7Bv6n7B52zi0Pjn65z/wvNe0Cbu/U6kUSpD5zySpBN8vlzrnz0l2LSDKpm0VEJAQU5iIiIaBuFhGREFDLXEQkBBTmIiIhoDAXEQkBhbmISAgozEVEQuD/AOdyqqRnh4nLAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 회귀 모델 평가\n",
        "model.evaluate(test_X, test_Y)\n",
        "#낮은 손실값"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VN6A94uOABhj",
        "outputId": "dcd42c2e-1fcc-4a73-ade8-4a86cbb9a8ee"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "4/4 [==============================] - 0s 3ms/step - loss: 0.2319\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.2319420874118805"
            ]
          },
          "metadata": {},
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#실제 주택 가격과 예측 주택 가격 시각화\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "pred_Y = model.predict(test_X)\n",
        "\n",
        "plt.figure(figsize=(8,8))\n",
        "plt.plot(test_Y, pred_Y, 'b.')\n",
        "plt.axis([min(test_Y), max(test_Y), min(test_Y), max(test_Y)])\n",
        "\n",
        "# y=x에 해당하는 대각선\n",
        "plt.plot([min(test_Y), max(test_Y)], [min(test_Y), max(test_Y)], ls=\"--\", c=\".3\")\n",
        "plt.xlabel('실제 주택 가격')\n",
        "plt.ylabel('예측 주택 가격')\n",
        "\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 573
        },
        "id": "hSLiGTocEHT3",
        "outputId": "5735c62b-e371-4d59-91ef-e2eba2736c0c"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/matplotlib/backends/backend_agg.py:214: RuntimeWarning: Glyph 8722 missing from current font.\n",
            "  font.set_text(s, 0.0, flags=flags)\n",
            "/usr/local/lib/python3.7/dist-packages/matplotlib/backends/backend_agg.py:183: RuntimeWarning: Glyph 8722 missing from current font.\n",
            "  font.set_text(s, 0, flags=flags)\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Root Mean Squared Error (제곱근 평균 제곱 오차)\n",
        "import math\n",
        "mse = tf.keras.losses.MeanSquaredError()\n",
        "math.sqrt(mse(train_Y, pred_Y).numpy())"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ThOlYcIS7vIo",
        "outputId": "1f7d029a-2711-496d-9e73-7fa5d89a9bfe"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "1.3287782296800208"
            ]
          },
          "metadata": {},
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Mean Absolute Error (평균 절대 오차)\n",
        "mae = tf.keras.losses.MeanAbsoluteError()\n",
        "mae(train_Y, pred_Y).numpy()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "yNHthVyC7yDg",
        "outputId": "9cce3002-9217-4654-fbbf-9adf006cfb9b"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "1.0029565"
            ]
          },
          "metadata": {},
          "execution_count": 13
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Mean Absolute Percentage Error (평균 절대 백분율 오차)\n",
        "mape = tf.keras.losses.MeanAbsolutePercentageError()\n",
        "mape(train_Y, pred_Y).numpy()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "8K15hM4X7zjA",
        "outputId": "697956c1-f5ba-4a84-c448-09ea1c00cc18"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "681.4621"
            ]
          },
          "metadata": {},
          "execution_count": 14
        }
      ]
    }
  ]
}