Abstract

We establish a sufficient condition ensuring strong Hellinger consistency of posterior distributions. We also prove a strong Hellinger consistency theorem for the pseudoposterior distributions based on the likelihood ratio with power 0 < alpha < 1, which are introduced by Walker and Hjort (2001). Our result is an extension of their theorem for alpha = 1/2

Keywords

Hellinger consistency; posterior distribution; nonparametric model

Published in

Research report (Centre of Biostochastics)
2008, number: 9, pages: 1-14
Publisher: Centre of Biostochastics, SLU

Authors' information