How it works...

In this recipe, we split the data using the train_test_split() function of the scikit-learn library. This function splits arrays or matrices into random train and testing subsets. Random division of input data into data sources for training and testing ensures that data distribution is similar for training and testing data sources. You choose this option when it is not necessary to preserve the order of the input data.