Weibull Racing Time-to-event Modeling and Analysis of Online Borrowers' Loan Payoff and Default

We propose Weibull delegate racing (WDR) to explicitly model surviving under competing events and to interpret how the covariates accelerate or decelerate the event time. It explains non-monotonic covariate effects by racing a potentially infinite number of sub-events, and consequently relaxes the ubiquitous proportional-hazards assumption which may be too restrictive. For inference, we develop a Gibbs-sampler-based MCMC algorithm along with maximum a posteriori estimations for big data applications. We analyze time to loan payoff and default on Prosper.com, demonstrating not only a distinguished performance of WDR, but also the value of standard and soft information.