A Bayesian significance test of change in the presence of a single outlier

Abstract

The Bayesian signicance test for a change in independent gaussian samples in the presence of single outlier is considered. The impact of an outlier on the performance of the Bayesian signicance test is studied.

Dans ce travail, nous considerons un test de signication Bayesien pour la detection de rupture dans un echantillon gaussien en presence d'une observation aberrante. L'impact d'une contamonation sur la performance du test est etudie.

Key words: Gaussian models,; Change point; HPD region sets; p-value; Outlier.