Abstract

ERRORS OF MISCLASSIFICATIONS ASSOCIATED WITH EDGEWORTHSERIES DISTRIBUTION (ESD)

Awogbemi Clement Adeyeye and Onyeagu Sidney Iheanyi

---

Abstract In this study, the discrimination and classification problems associated with the persistent non-normal distribution have been studied. Sampling from non-normal distribution is assessed through the distribution of errors of misclassification in respect of Edgeworth Series Distribution (ESD). The effects of applying a normal distribution (ND) classificatory rule when the distribution is ESD by empirical approach is examined by comparing the errors of misclassification for ESD with ND using small sample sizes at every level of skewness factor. The expected probability of misclassification for ESD was generated through the characteristic function and conditional expectation approach. The asymptotic distribution of the error of misclassification for ESD was also derived through Taylor?s expansion which was used to determine the approximate mean and variance for ESD. Keywords: Errors of Misclassification, Edgeworth Series Distribution, Skewness Factor, Classificatory Rule, Optimum Probability of Misclassification

PDF