survivalmodels

Models for Survival Analysis

Implementations of classical and machine learning models for survival analysis, including deep neural networks via 'keras' and 'tensorflow'. Each model includes a separated fit and predict interface with consistent prediction types for predicting risk, survival probabilities, or survival distributions with 'distr6' <https://CRAN.R-project.org/package=distr6>. Models are either implemented from 'Python' via 'reticulate' <https://CRAN.R-project.org/package=reticulate>, from code in GitHub packages, or novel implementations using 'Rcpp' <https://CRAN.R-project.org/package=Rcpp>. Novel machine learning survival models wil be included in the package in near-future updates. Neural networks are implemented from the 'Python' package 'pycox' <https://github.com/havakv/pycox> and are detailed by Kvamme et al. (2019) <https://jmlr.org/papers/v20/18-424.html>. The 'Akritas' estimator is defined in Akritas (1994) <doi:10.1214/aos/1176325630>. 'DNNSurv' is defined in Zhao and Feng (2020) <arXiv:1908.02337>.

Functions (26)