DeepLearning - 타이타닉 생존자 예측

'''타이타닉셋을 가져와서, 딥러닝으로 Survived 예측하기'''

md = keras.Sequential()

md.add(Dense(64, activation="sigmoid"))

md.add(Dense(32, activation="sigmoid"))

md.add(Dense(1, activation="sigmoid"))

# 컴파일

md.compile(optimizer='adam', loss='binary_crossentropy', metrics=['binary_accuracy'])

md.fit(X_train,y_train,epochs=5000)

kg_up = df_kg.copy()

pred_ = md.predict(df_kg).reshape(df_kg.shape[0])

pred_ = np.where(pred_ >0.5,1,0)

kg_up['Survived'] = pred_

kg_up.Survived.to_csv('kaggle_deep_clf.csv')

 

참고 : 점수 낮게 나옴 ! 

0.71~