Premium
Evaluation of misclassification errors
Author(s) -
Srivastava M. S.
Publication year - 1973
Publication title -
canadian journal of statistics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.804
H-Index - 51
eISSN - 1708-945X
pISSN - 0319-5724
DOI - 10.2307/3314646
Subject(s) - statistic , covariance , covariance matrix , statistics , computer science , multivariate statistics , matrix (chemical analysis) , expression (computer science) , mathematics , algorithm , materials science , composite material , programming language
In this paper the problem of classifying an individual into one of two multivariate normal populations with common covariance matrix is considered. An attempt is made to provide some idea about the errors of misclassification for the Welch‐Anderson statistic. It is shown that in most cases the usual asymptotic expression (the one when all the parameters are known) is an underestimate of the actual error committed. In Part II, an alternative classification procedure, known as a step‐down procedure, is proposed. In this the errors of misclassification are easily computable.