train_with_sklearn_keras.py

SEED = 7

import numpy as np
np.random.seed(SEED)

from keras.models import Sequential
from keras.layers import Dense, Dropout

def create_model(input_dim, hidden_units, kernel_initializer, activation, dropout_rate, loss, optimizer, metrics):
  model = Sequential()

  for index, hidden_unit in enumerate(hidden_units):
    if index == 0:
      model.add(Dense(hidden_unit, kernel_initializer=kernel_initializer, activation=activation, input_dim=input_dim))
    else:
      model.add(Dense(hidden_unit, kernel_initializer=kernel_initializer, activation=activation))
    model.add(Dropout(dropout_rate))

  model.add(Dense(1, kernel_initializer=kernel_initializer, activation='sigmoid'))

  model.compile(loss=loss, optimizer=optimizer, metrics=metrics)
  return model

def main():

  from keras.wrappers.scikit_learn import KerasClassifier

  model = KerasClassifier(build_fn=create_model
                         ,input_dim=686
                         ,hidden_units=[256, 256, 256]
                         ,kernel_initializer='uniform'
                         ,activation='relu'
                         ,dropout_rate=0.4
                         ,loss='binary_crossentropy'
                         ,optimizer='adam'
                         ,metrics=['accuracy']
                         ,epochs=50
                         ,batch_size=128
                         )

  import hppi

  hppids = hppi.read_data_sets("data/02-ct-bin", one_hot=False)
  X = hppids.datas
  Y = hppids.labels

  from sklearn.model_selection import StratifiedKFold, cross_val_score

  kfold = StratifiedKFold(n_splits=10, shuffle=True, random_state=SEED)
  results = cross_val_score(model, X, Y, cv=kfold)

  print(np.average(results))

if __name__ == "__main__":
  main()