New algorithms for finding approximate frequent item sets