Weighted Csiszár-Kullback-Pinsker inequalities and applications to transportation inequalities (Q816460): Difference between revisions

The Csiszár-Kullback-Pinsker inequality compares the total variation of the difference of two probability measures and the Kullback information of one measure with respect to another one. The authors strengthen this inequality by allowing weights in the total variation norm. The class of admissible weights depends on the decay properties of the reference probability measure. This result is used to derive transportation inequalities involving Wasserstein distances for various exponents. An application to transportation inequalities for random dynamical systems is given.