Word-Map Systems for Content-Based Document Classification

The main purpose of this paper is the classification of documents in terms of their content. Two systems are presented here that share a two-level architecture that include 1) a word map created via unsupervised learning that functions as a document-representation module and 2) a supervised multilayer-perceptron-based classifier. Two approaches to create word maps are presented and compared; these are based on hidden Markov models (HMMs) and the self-organizing map. A series of experiments is performed on several datasets of text-only documents, which are written in either Greek or in English. A comparison with established methods, such as the support-vector machine (SVM), illustrates the effectiveness of the proposed systems.