Kernel Density Machines

Andrea Della Vecchia, Damir Filipovic, Paul Schneider

We introduce kernel density machines (KDM), an agnostic kernel-based framework for learning the Radon-Nikodym derivative (density) between probability measures under minimal assump…