Hands-On Machine Learning for Algorithmic Trading

How to use logistic regression for prediction

The lasso L₁ penalty and the ridge L₂ penalty can both be used with logistic regression. They have the same shrinkage effect as we have just discussed, and the lasso can again be used for variable selection with any linear regression model.

Just as with linear regression, it is important to standardize the input variables as the regularized models are scale sensitive. The regularization hyperparameter also requires tuning using cross-validation as in the linear regression case.