Deep Learning with Tensorflow

A sample on how to use Scikit and Tensorflow to predict survival on the Titanic. The data is obtained from kaggle.

I published my code/results on Github as part of my Data Science hype presentation.

Also check out my small Modern Portfolio Theory with R project!

Sample Result

INFO:tensorflow:Step 101: loss = 0.568274
INFO:tensorflow:Step 201: loss = 0.554627
INFO:tensorflow:Step 301: loss = 0.546914
INFO:tensorflow:Step 401: loss = 0.540043
INFO:tensorflow:Step 501: loss = 0.534155
INFO:tensorflow:Step 601: loss = 0.529909
INFO:tensorflow:Step 701: loss = 0.523719
INFO:tensorflow:Step 801: loss = 0.520302
INFO:tensorflow:Step 901: loss = 0.518217
INFO:tensorflow:Loss for final step: 0.511472.

Decision Tree Prediction Score: 0.7027027027027027
DNN Prediction Score: 0.6756756756756757
Random Forest Prediction Score: 0.6486486486486487