Two-group classification procedures with multivariate binary variables

Over the years, a considerable body of research has accumulated on classification analysis, with its usefulness demonstrated in various fields, including engineering, medical and social sciences, economics, marketing, finance, education and management. In this book, eight classification rules for binary variables were considered. A simulation experiment was conducted to compare the performance of all these rules. The eight classification procedures are discussed and evaluated at 118 configurations of the sampling experiments. From the analysis carried out, the results obtained ranked the procedures as follows: Optimal, Linear discriminant, Maximum likelihood, Predictive, Dillon-Goldstein, Full multinomial, Likelihood ratio and Nearest neighbour rule. Whereas for three and four variables, the maximum likelihood rule is preferred and for five variables optimal rule is preferred in terms of minimizing the expected actual error rate. The overall best method is the optimal classification rule. This book is intended to be useful to professionals in medical, Education, Social sciences, finance or anyone studying probability and multivariate data statistics at sub-degree and degree levels.

Autorentext

Egbo Ikechukwu is a principal lecturer in Mathematics and Statistics, Alvan Ikoku Federal College of Education, Owerri-Nigeria. He possesses B.Sc. in Mathematics from University of Nigeria, Nsukka, PGD and M.Sc. in Statistics from Nnamdi Azikwe University, Awka-Nigeria as well as a PhD in Statistics from Imo State University, Owerri-Nigeria.