6.7 - Final thoughts
Regularized regression provides many great benefits over traditional GLMs when applied to large data sets with lots of features.
It provides an option for handling the \(n\) > \(p\) (i.e., more observations than features), minimize the impact of multicollinearity, and can perform automated feature selection.
It also has relatively few hyperparameters (i.e., makes it easier to tune), is computationally and memory efficient, compared to other algorithms.