Hands-On Machine Learning for Algorithmic Trading

Ridge regression using sklearn

For the ridge regression, we need to tune the regularization parameter with the keyword alpha that corresponds to the λ we used previously. We will try 21 values from 10^-5 to 10⁵ in logarithmic steps.

The scale sensitivity of the ridge penalty requires us to standardize the inputs using the StandardScaler. Note that we always learn the mean and the standard deviation from the training set using the .fit_transform() method and then apply these learned parameters to the test set using the .transform() method.