In this version, we also perform a verification of the selected features ( correlation ) with a PCA and we reconsider our decision after this comparison.

PhD Program in Statistics DISIA , University of Florence

This analysis compares the standard Cox proportional hazards model with a flexible parametric Royston–Parmar model using the lung cancer dataset from the survival package. We fit both models using age and sex as covariates, extract the cumulative baseline hazard from the Cox model, estimate the hazard function from the Royston–Parmar model, and visualize both on the same graph for comparison. The study highlights differences between stepwise and smooth hazard estimates and demonstrates the interpretability of flexible parametric survival models.

NOAA Storm Data Analysis: Population & Economic Impacts