This assignment applies three modern regression techniques—Ridge, LASSO, and Partial Least Squares (PLS)—to a real dataset with 10+ predictors and a continuous outcome. You will review the core ideas and formulas behind each method, then use R to fit the models, tune hyper-parameters via cross-validation, and generate coefficient tables, selection metrics (e.g., RMSE, MAE, R²), and diagnostic plots. The analysis compares which variables LASSO retains or removes, how Ridge and PLS address multicollinearity differently, and which approach yields the strongest predictive performance.

Mathematical modelling and Simulations Assignment One