| Article ID: | iaor2007442 |
| Country: | United States |
| Volume: | 17 |
| Issue: | 1 |
| Start Page Number: | 25 |
| End Page Number: | 31 |
| Publication Date: | Dec 2005 |
| Journal: | INFORMS Journal On Computing |
| Authors: | Street W. Nick |
| Keywords: | artificial intelligence |
Induction of decision trees is a popular and effective method for solving classification problems in data-mining applications. This paper presents a new algorithm for multi-category decision tree induction based on nonlinear programming. This algorithm, termed OC-SEP (Oblique Category SEParation), combines the advantages of several other methods and shows improved generalization performance on a collection of real-world data sets.