Text Processing Adaptive Resonance Theory Neural Network

Kreesuradej, Worapoj; Chantasut, Norraseth

Text Processing Adaptive Resonance Theory Neural Network

Author(s)

Kreesuradej, Worapoj

Chantasut, Norraseth

Date Issued

December 1, 2002

Type

Conference Paper

Abstract

This paper proposes a Text Processing Adaptive Resonance Theory Neural Network for document clustering. Unlike the conventional clustering algorithms, a Text Processing Adaptive Resonance Theory Neural Network works directly on textual information without transforming text data into a numerical value. The main contribution of this paper is to show how to adapt the concepts of ART1 clustering on a data set, which has a qualitative feature values. The Text Processing Adaptive Resonance Theory Neural Network utilizes of the concept of similarity measure for symbolic objects, which is different from the conventional similarity measure for objects whose feature values are numerical values. The proposed neural network assigns cluster labels to the objects.

Citation

Intelligent Engineering Systems Through Artificial Neural Networks, 12, 625-630, 2002