论文信息 - The asymptotic distributions of the conditional error rate and risk in discriminant analysis

The asymptotic distributions of the conditional error rate and risk in discriminant analysis

SUMMARY The distributions of the conditional error rate and risk associated with Anderson's classification statistic in the context of the two-population discrimination problem with different means and common covariance matrix are studied. For the most general case where all three population parameters are unknown, the distributions are able to be approximated in the form of asymptotic expansions of degrees higher than previously available. Some key word8: Anderson's classification statistic; Asymptotic distribution; Conditional error rate and risk in discriminant analysis.

G. McLachlan

[1] T. W. Anderson. Classification by multivariate analysis , 1951 .

[2] S. John. Errors in Discrimination , 1961 .

[3] M. Okamoto. An Asymptotic Expansion for the Distribution of the Linear Discriminant Function , 1963 .

[4] M. Hills. Allocation Rules and Their Error Rates , 1966 .

[5] O. J. Dunn. Some Expected Values for Probabilities of Correct Classification in Discriminant Analysis , 1971 .

[6] G. McLachlan. AN ASYMPTOTIC EXPANSION FOR THE VARIANCE OF THE ERRORS OF MISCLASSIFICATION OF THE LINEAR DISCRIMINANT FUNCTION1 , 1972 .

[7] G. McLachlan. Asymptotic Results for Discriminant Analysis When the Initial Samples are Misclassified , 1972 .