Paper:
Language: English:

Improving Text Categorization by Multicriteria Feature Selection

Son Doan^*, and Susumu Horiguchi^**

*Graduate School of Information Science, Japan Advanced Institute of Science and Technology, 1-1 Asahidai, Nomi, Ishikawa 923-1292, Japan
**Graduate School of Information Science, Tohoku University, 6-3-09 Aoba, Sendai 980-8579, Japan

Received: October 28, 2004

Accepted: May 9, 2005

Keywords: feature selection, text categorization, machine learning, text mining, text representation

Journal ref: Journal of Advanced Computational Intelligence and Intelligent Informatics, Vol.9, No.5 pp. 570-575, 2005

Abstract

Text categorization involves assigning a natural language document to one or more predefined classes. One of the most interesting issues is feature selection. We propose an approach using multicriteria ranking of eatures, a new procedure for feature selection, and apply these to text categorization. Experimental results dealing with Reuters-21578 and 20Newsgroups benchmark data and the naive Bayes algorithm show that our proposal outperforms conventional feature selection in text categorization performance.

[Notice]
* "Preview" is the first 2 pages of the article. You don't need the registration.
* To read the PDF file you will then need to download and install the Adobe Reader.
Adobe Reader is free and available for download here: