Weibull Racing Survival Analysis for Competing Events and a Study of Loan Payoff and Default

We propose Bayesian nonparametric Weibull delegate racing (WDR) to explicitly model surviving under competing events and to interpret how the covariates accelerate or decelerate the event times. WDR explains non-monotonic covariate effects by racing a potentially infinite number of sub-events, relaxing the ubiquitous proportional-hazards assumption which may be too restrictive. WDR can handle different types of censoring and missing event times or types. For inference, we develop a Gibbs-sampler-based MCMC algorithm along with a maximum a posteriori estimation for big data applications. We use synthetic data analysis to demonstrate the flexibility and parsimonious nonlinearity of WDR. We also use a data set of time to loan payoff and default from Prosper.com to showcase the interpretability.