. . . "The combination of multiple features or views when representing documents or other kinds of objects usually leads to improved results in classification (and retrieval) tasks." .