A Sensitive Attribute based Clustering Method for kanonymization

In medical organizations large amount of personal data are collected and analyzed by the data miner or researcher, for further perusal. However, the data collected may contain sensitive information such as specific disease of a patient and should be kept confidential. Hence, the analysis of such data must ensure due checks that ensure protection against threats to the individual privacy. In this context, greater emphasis has now been given to the privacy preservation algorithms in data mining research. One of the approaches is anonymization approach that is able to protect private information; however, valuable information can be lost. Therefore, the main challenge is how to minimize the information loss during an anonymization process. The proposed method is grouping similar data together based on sensitive attribute and then anonymizes them. Our experimental results show the proposed method offers better outcomes with respect to information loss and execution time.