Training Data shape: (1894676, 228)
Test Data shape:      (933200, 228)
AutoEncoder(batch_size=1024, contamination=0.1966,
      dropout_rate=0.2, epochs=100, hidden_activation='relu',
      hidden_neurons=[128, 64, 32, 64, 128], l2_regularizer=0.1,
      loss='binary_crossentropy', optimizer='adam',
      output_activation='sigmoid', preprocessing=False, random_state=42,
      validation_size=0.1, verbose=1)

Epoch 100/100
1705208/1705208 [==============================] - 72s 42us/step - loss: -2.1918e-04 - val_loss: 0.0103

AUC: 0.701
Brier: 0.193
Precision: 0.337
Recall: 0.337
F1: 0.337