Spatio-temporal Bayesian On-line Changepoint Detection with Model Selection
Jeremias Knoblauch (University of Warwick)
Friday 6th July, 2018 15:00-16:00 Maths 311B
Bayesian On-line Changepoint Detection is extended to on-line model selection and non-stationary spatio-temporal processes. We propose spatially structured Vector Autoregressions (VARs) for modelling the process between changepoints (CPs) and give an asymptotic upper bound on the approximation error of such models. The resulting inference procedure performs prediction, model selection and CP detection in multivariate data on-line. Its time complexity is linear and its space complexity constant, and thus it is two orders of magnitudes faster than its closest competitor. In addition, its additional flexibility in modelling changing dynamics and dependency structures allow it to outperform the state of the art on multivariate data.