Bayesian Inference for Change Points in Dynamical Systems with Reusable States - a Chinese Restaurant Process Approach
Proceedings of the Fifteenth International Conference on Artificial Intelligence and Statistics, PMLR 22:1117-1124, 2012.
We study a model of a stochastic process with unobserved parameters which suddenly change at random times. The possible parameter values are assumed to be from a finite but unknown set. Using a Chinese restaurant process prior over parameters we develop an efficient MCMC procedure for Bayesian inference. We demonstrate the significance of our approach with an application to systems biology data.