Asymptotic Expansions for Errors of Misclassification: Nonnormal Situations

A unified development of the expansions available for the asymptotic distribution of the errors of misclassification is given in the nonnormal situations. Several examples are presented.A detailed study is undertaken in the case of outliers in sampling from the normal distribution. It is Shown that outliers with large variances lead to inflation in the errors of misclassification.