train_model.py

import pandas as pd

import keras
from keras.callbacks import TensorBoard
from keras.models import Sequential
from keras.layers import *

RUN_NAME = "Run 1 with 50 nodes"

training_data_df = pd.read_csv("sales_data_training_scaled.csv")

X = training_data_df.drop('total_earnings', axis=1).values
Y = training_data_df[['total_earnings']].values

# Define the model
model = Sequential()
model.add(Dense(50, input_dim=9, activation='relu', name='layer_1'))
model.add(Dense(100, activation='relu', name='layer_2'))
model.add(Dense(50, activation='relu', name='layer_3'))
model.add(Dense(1, activation='linear', name='output_layer'))
model.compile(loss='mean_squared_error', optimizer='adam')

# Create a TensorBoard logger
tensorboard = TensorBoard(
    log_dir='logs/{}'.format(RUN_NAME),
    histogram_freq=0,
    write_graph=True,
)

# Train the model
model.fit(
    X,
    Y,
    epochs=50,
    shuffle=True,
    verbose=2,
    callbacks=[tensorboard],
)

# Load the separate test data set
test_data_df = pd.read_csv("sales_data_test_scaled.csv")

X_test = test_data_df.drop('total_earnings', axis=1).values
Y_test = test_data_df[['total_earnings']].values

test_error_rate = model.evaluate(X_test, Y_test, verbose=0)
print("The mean squared error (MSE) for the test data set is: {}".format(test_error_rate))

# Save the model to disk
model.save("trained_model.h5")
print("Model saved to disk.")