 Risk consistency

A sequence of classifiers $h_n$ is risk-consistent when its classification risk converges to the Bayes risk as the training sample size tends to infinity.