The Multivariate Bernoulli detector: Change point estimation in discrete survival analysis

• DOI: 10.48550/arxiv.2308.10583
• Publication Date: 2023-01-01
• Authors: Willem van den Boom; Maria De Iorio; Fang Qian; Alessandra Guglielmi
• Keywords: FOS: Computer and information sciences; Models, Statistical; Biometry; Bayes Theorem; Survival Analysis; 01 natural sciences; Markov Chains; Methodology (stat.ME); Multivariate Analysis; Humans; Computer Simulation; 0101 mathematics; Monte Carlo Method; Statistics - Methodology; Algorithms

Time-to-event data are often recorded on a discrete scale with multiple, competing risks as potential causes for the event. In this context, application of continuous survival analysis methods single risk suffer from biased estimation. Therefore, we propose Multivariate Bernoulli detector times involving multivariate change point model cause-specific baseline hazards. Through prior number points and their location, impose dependence between across risks, well allowing data-driven learning number. Then, conditionally these points, is used to infer which involved. Focus posterior inference hazard rates risks. Such present due subject-specific changes time that affect all Full performed through tailored local-global Markov chain Monte Carlo (MCMC) algorithm, exploits augmentation trick MCMC updates non-conjugate Bayesian nonparametric methods. We illustrate our in simulations prostate cancer data, comparing its performance existing approaches.