Kernel based method for the $k$-sample problem

Abstract

In this paper we deal with the problem of testing for the equality of $k$ probability distributions defined on $(\mathcal{X},\mathcal{B})$, where $\mathcal{X}$ is a metric space and $\mathcal{B}$ is the corresponding Borel $\sigma$-field. We introduce a test statistic based on reproducing kernel Hilbert space embeddings and derive its asymptotic distribution under the null hypothesis. Simulations show that the introduced procedure outperforms known methods.

arXiv comment30 pages, 4 figures

References in corpus (1)

• Testing for Homogeneity with Kernel Fisher Discriminant Analysis

Cited by in corpus (2)

• Comparing a Large Number of Multivariate Distributions
• Energy distance and kernel mean embeddings for two-sample survival testing