A text processing Kohonen neural network

Abstract

This paper proposes a text processing Kohonen neural network. This algorithm works directly on textual information without mapping documents into some representation that has quantitative features. The input level of the proposed neural network can directly receive a qualitative value without mapping the qualitative value into numerical value. Then, based on Kohonen self organize feature maps and the concepts of dissimilarity measure for symbolic objects, the proposed neural network assigns cluster labels to the objects.