Simultaneous Clustering and Dynamic Keyword Weighting for Text Documents

Abstract

In this chapter, we propose a new approach to unsupervised text document categorization based on a coupled process of clustering and cluster-dependent keyword weighting. The proposed algorithm is based on the K-Means clustering algorithm. Hence it is computationally and implementationally simple. Moreover, it learns a different set of keyword weights for each cluster. This means that, as a by-product of the clustering process, each document cluster will be characterized by a possibly different set of keywords. The cluster-dependent keyword weights have two advantages: they help in partitioning the document collection into more meaningful categories; and they can be used to automatically generate a compact description of each cluster in terms of not only the attribute values,but also their relevance. In particular, for the case of text data, this approach can be used to automatically annotate the documents. We also extend the proposed approach to handle the inherent fuzziness in text documents, by automatically generating fuzzy or soft labels instead of hard all-or-nothing categorization. This means that a text document can belong to several categories with different degrees. The proposed approach can handle noise documents elegantly by automatically designating one or two noise magnet clusters that grab most outliers away from the other clusters. The performance of the proposed algorithm is illustrated by using it to cluster real text document collections.