사업성과
논문

논문

Home / 사업성과

Wasserstein upper bounds of the total variation for smooth densities

Jounal

STATISTICS & PROBABILITY LETTERS

Vol

163

SCI(E) 구분

Author

Minwoo Chae, Stephen G. Walker

CONTENT

The total variation distance between probability measures cannot be bounded by the Wasserstein metric in general. If we consider sufficiently smooth probability densities, however, it is possible to bound the total variation by a power of the Wasserstein distance. We provide a sharp upper bound which depends on the Sobolev norms of the densities involved. (C) 2020 Elsevier B.V. All rights reserved.