A classification rule based on the minimum Kolmogorov distance for classifying an individual into one of two univariate populations is proposed. Some almost sure convergence theorems are derived and then utilized to study the limiting distributions of the conditional probabilities of correct classification. It is shown that the asymptotic distributions of the conditional probabilities of correct classification are normal under some mild conditions.
關聯:
Communications in Statistics: Theory and Methods 30(2), pp.319-330