Document categorization and retrieval using semantic microfeatures and growing cell structures

Presents an approach to document categorization and retrieval using growing cell structures. Semantic microfeatures together with other information are adopted for document representation. The main advantage over traditional information retrieval systems which adopt index terms to index and retrieve documents is that the new approach considers the semantic relationships among documents. The paper describes how to construct document microfeature weight vectors and use growing cell structures clustering them. Some experimental results are also presented.