Anonymizing but Deteriorating Location Databases

The tremendous development of location-based services and mobile devices has led to an increase in location databases. Through the data mining process, valuable information can be discovered from such location databases. However, the malicious data miner or attackers may also extract private and sensitive information about the user, and this can create threats against the user location privacy. Therefore, location privacy protection becomes a key factor to the success in privacy protection for the users of location-based services. In this paper, we propose a novel approach as well as an algorithm to guarantee k-anonymity in a location database. The algorithm will maintain the association rules that have significance for the data mining process. Moreover, there may appear new significant association rules created after anonymization, they maybe affect the data mining result. Therefore, the algorithm also considers excluding new significant association rules that are created during the run of the algorithm. Theoretical analyses and experimental results with real-world datasets will confirm the practical value of our newly proposed approach.