jeroenhoogland/regsurv: Regularized parametric survival modeling

Provides functions to develop regularized flexible hazard and cumulative hazard models on right-censored survival data. Implemented regularization options include lasso, ridge, elastic net and groups lasso penalties, which may be combined within a single model. Flexible refers to spline-based baseline hazard or cumulative baseline hazard specification. The default is to model proportional covariate effects, but the regsurv package allows for easy incorporation of time varying effects.

install.packages("remotes")
remotes::install_github("jeroenhoogland/regsurv")