Delegating Classifiers for Automatic Text Categorization

Abstract

This paper presents delegating classifiers for text categorization. Delegating classifiers for text categorization avoid some of the problems in ensemble methods, in particular the loss of comprehensibility and the use of excessive computational resources. Delegation is a technique with a very good trade off between the quality of models and the resources needed to obtain them. Delegation is a serial multi classifier method, not only in learning but also in classification.