We consider data that informe about the level of prostate-specific antigen and a number of clinical measures on 97 men. Of interest in the study is the relationship between the prostate-specific antigen (psa) and the clinical measures standing for predictor variables. The data frame consists in 97 observations on 8 predictor variables and the response variable (lpsa).
The prediction modelling is achieved with regression models. We first apply the standard linear regression as a baseline, then we consider regularized regression models:
- Ridge regression
- Lasso
- Elastic Net
We study the influence of regularization on the model.