Оценивание смещения эмпирического риска для линейных классификаторов

An empirical risk bias in classification problem is researched. Statistical modeling performed shows that the risk bias dependence on decision class capacity appears to be the same both for multinomial (discrete) case and for linear classifier. This result ensures that universal scaling of Vapnik-Chervonenkis bias estimations may be available since such scaling was obtained for a discrete case. To prove using an empirical risk estimator a comparison of it’s volatility versus volatility of leave-one-out estimator is also performed.