Linear Regression Using Python

This week, we consider regression having more advanced aspects:

use of regularisation when some predictors are correlated and hence too similar to each other
use of cross-validation for
model building: deciding the polynomial degree to predict a target
hyperparameter tuning: trying to find the best value of the regularisation parameter lambda

The first example relates to fitting a polynomial to data, with or without regularisation, using cross-validation to derive a test score that can be compared for different polynomial degrees.

The second example relates to the analysis of Advertising data by media channel. We also covered this example scenario in class.