Python Feature Engineering Cookbook

How it works...

In this recipe, we performed feature hashing using the Category Encoders package. We first loaded the dataset and divided it into train and test sets using train_test_split() from scikit-learn. Next, we set up HashingEncoder() to encode the A7 variable into four components. With the fit() method, HashingEncoder() created a mapping from the category to a set of binary vectors and with the transform() method, the encoder encoded the A7 variable in both the train and test sets.

With one-hot encoding, we would have needed nine binary variables to encode A7. With feature hashing, we can represent the variable in a smaller feature space, that is, with fewer derived features. The number of features is determined by the user.