keras_boston_example.py

import numpy as np
import sklearn.datasets as datasets
from keras import layers, models, utils


def main():
    # Loading dataset
    boston_dataset = datasets.load_boston()

    # Defining inputs and outputs of the neural network
    X = np.array(boston_dataset['data'])
    y = np.array(boston_dataset['target'])

    # Defining model with 2 layers
    # One containing 100 neurons and a output layer
    model = models.Sequential()
    model.add(layers.Dense(100, input_shape=(13, )))
    model.add(layers.Dense(1))

    # Defining the model parameters
    model.compile(
        optimizer='adam',
        loss='mean_squared_error',
        metrics=['accuracy', 'mae'])

    # Training the model
    model.fit(X, y, epochs=1000, validation_split=0.1, verbose=2)


if __name__ == '__main__':
    main()