CMStatistics 2018: Start Registration
View Submission - CMStatistics
Title: Stein kernels and information Authors:  Yvik Swan - Universite de Liege (Belgium) [presenting]
Gesine Reinert - Oxford University (United Kingdom)
Marie Ernst - University of Liege (Belgium)
Abstract: A general concept of Stein kernels is presented and the corresponding covariance identities are developed. We propose a notion of Stein-Fisher information. Among many possible applications, we extract a general tool for goodness-of-fit testing based on recent works concerning ``kernelized stein discrepancy for goodness-of-fit tests''. We insist mainly on the discrete framework for the examples.