Improper linear discriminant classifiers

Improper linear discriminant classifiers

0.00 Avg rating0 Votes
Article ID: iaor1993818
Country: Netherlands
Volume: 50
Issue: 2
Start Page Number: 188
End Page Number: 198
Publication Date: Jan 1991
Journal: European Journal of Operational Research
Authors:
Abstract:

Mathematical programming formulations for determining linear discriminant classifiers have received considerable interest in recent years. Although their performance rivals classical methods and requires fewer assumptions, many of the formulations exhibit undesirable properties. This paper draws attention to two additional problems that the mathematical programming formulations face. A solution for each type of problem is presented.

Reviews

Required fields are marked *. Your email address will not be published.