The thesis reports the studies performed to improve the ROCK clustering algorithm together with its applications in the anonymization and anomaly detection fields. Various changes to the original algorithm are introduced and the algorithm itself is employed in a novel environment to achieve better anonymization results.