Identifying groups in data using cluster K-means