Article ID: | iaor1997713 |
Country: | United States |
Volume: | 14 |
Issue: | 3/4 |
Start Page Number: | 267 |
End Page Number: | 287 |
Publication Date: | Jul 1994 |
Journal: | American Journal of Mathematical and Management Sciences |
Authors: | Xiao Baichun |
Keywords: | decision theory, statistics: general |
Discriminant analysis is an important approach used to solve many practical problems in various fields. A number of mathematical programming models have been developed for this analysis. While some drawbacks with linear programming models have been discussed, little has been done on nonlinear programming (NLP) models due to their complexity. This paper examines the significance of multiple and unacceptable solutions in NLP discriminant analysis, identifies the conditions under which these solutions are generated, and proposes a regularization method to eliminate multiple solutions.