Automatic text classification of English newswire articles based on statistical classification techniques

Abstract

The basic process of automatic text classification is learning a classification scheme from training examples and then using it to classify unseen textual documents. It is essentially the same as the process of graphic or character pattern recognition. Thus, the pattern recognition approaches can be used for automatic text categorization. In this research several statistical classification techniques that include Euclidean distance, various similarity measures, linear discriminant function, projection distance, modified projection distance, and SVM, have been used for automatic text classification. Principal component analysis was used to reduce the dimensionality of the feature vector. Comparative experiments have been conducted using the Reuters-21578 test collection of English newswire articles. The results illustrate that the overall efficiency of modified projection distance is better than the other methods and that principal component analysis is suitable for reducing the dimensionality of the text features. © 2005 Wiley Periodicals, Inc. Electr Eng Jpn, 152(1): 50–60, 2005; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/eej.20108