AN EMPIRICAL BAYESIAN ANALYSIS OF SIMULTANEOUS CHANGEPOINTS IN MULTIPLE DATA SEQUENCES By

Motivated by applications in genomics, finance, and biomolecular simulation, we introduce a Bayesian framework for modeling changepoints that tend to co-occur across multiple related data sequences. We infer the locations and sequence memberships of changepoints in our hierarchical model by developing efficient Markov chain Monte Carlo sampling and posterior mode finding algorithms based on dynamic programming recursions. We further propose an empirical Bayesian Monte Carlo expectation-maximization procedure for estimating unknown prior parameters from data. The resulting framework accommodates a broad range of data and changepoint types, including real-valued sequences with changing mean or variance and sequences of counts or binary observations. We demonstrate on simulated data that our changepoint estimation accuracy is competitive with the best methods in the literature, and we apply our methodology to the discovery of DNA copy number variations in cancer cell lines and the analysis of historical price volatility in U.S. stocks.