Hands-On Machine Learning for Algorithmic Trading

How to prepare the data

We can load it in Keras out of the box:

from keras.datasets import mnist(X_train, y_train), (X_test, y_test) = mnist.load_data()X_train.shape, X_test.shape((60000, 28, 28), (10000, 28, 28))

The following screenshot shows the first ten images in the dataset, and highlights significant variation among instances of the same digit. On the right, it shows how the pixel values for an individual image range from 0 to 255:

We rescale the pixel values to the range [0, 1] in order to normalize the training data, facilitate the backpropagation process, and convert the data to 32 bit floats that reduce memory requirements and computational ...