Experimental Comparison of Semi-parametric, Parametric, and Machine Learning Models for Time-to-Event Analysis Through the Concordance Index

In this paper, we make an experimental comparison of semi-parametric (Cox model, Aalen additive model), parametric (Weibull AFT model), and machine learning methods (Random Survival Forest, Gradient Boosting, DeepSurv) through the concordance index on two different datasets (PBC and GBCSG). We present two comparisons: one with the default hyperparameters of these methods and one with the best hyperparameters found by randomized search.