| Article ID: | iaor19972535 |
| Country: | United Kingdom |
| Volume: | 24 |
| Issue: | 42 |
| Start Page Number: | 179 |
| End Page Number: | 191 |
| Publication Date: | Feb 1997 |
| Journal: | Computers and Operations Research |
| Authors: | Loucopoulos Constantine |
| Keywords: | statistics: general |
A new mathematical programming model for the three-group classification problem is presented. This model is shown to be computationally more efficient than the General Single Function Classification model for a variety of data configurations. Theoretical characteristics of the classificatory performance of the two models are also investigated.