Reliability Estimation of Classification Algorithms. II. Point Bayesian Estimates

This is the second of three articles that deal with estimating the reliability of classification algorithms. It examines the statistical methods used for point estimation of the reliability of classifiers in the Bayesian approach. The case of unequally weighted precedents is considered. The estimation results are applicable to the small-sample case, which is of considerable importance in practice.__________Translated from Prikladnaya Matematika i Informatika, No. 16, pp. 90–104, 2004.