×

Bayesian change-point estimation in the presence of a single outlier. (English) Zbl 1258.62025

Summary: Bayesian estimation of change-points in independent Gaussian samples is considered. The impact of an outlier on the performance of the Bayesian procedure is studied and the posterior density under contamination is given. We show that, in the case of small samples, an outlier has a significant impact on the estimation of the change-point. However, if the sample size is reasonably high, the estimation performs well.

MSC:

62F15 Bayesian inference
62F35 Robustness and adaptive procedures (parametric inference)