A feature mining based approach for the classification of text documents into disjoint classes