Deep learning time

Finally, we will use deep learning to solve the issue and look for the accuracy of the results. We will take advantage of the keras package to use the Sequential and Dense models, and the KerasClassifier packages, as shown in the following code:

from keras.models import Sequential
from keras.layers import Dense
from keras.wrappers.scikit_learn import KerasClassifier

We change our function to have multiple hidden layers in our network:

def network_builder(hidden_dimensions, input_dim):
     # create model
     model = Sequential()
     model.add(Dense(hidden_dimensions[0], input_dim=input_dim,           kernel_initializer='normal', activation='relu'))
     # add multiple hidden layers
     for dimension in hidden_dimensions[1:]:
         model.add(Dense(dimension, kernel_initializer='normal', activation='relu'))
         model.add...