Efficiency concerns in Privacy Preserving Association Rule Mining-Optimization of algorithm

Shipra Agrawal

Abstract:An interesting new direction for data mining research is the development of techniques that incorporate privacy concerns. Being an emerging field, major concentration so far has been on defining the metrics of privacy and establishing the technical feasibility of development of accurate models about aggregated data while meeting the goals of privacy. Thus the goal of the research in privacy preserving data mining focused on meeting the two ends – privacy and accuracy and the third dimension – efficiency has not been much investigated. Existence of a large number of algorithms proposed for fast mining of association rules speaks for the importance of efficiency in association rule mining. It is evident that privacy preserving in association rule mining can’t become a reality if it costs excessively in terms of running time. This project attempts to incorporate efficiency into the privacy preserving association rule mining algorithm MASK wile ensuring that it does not incur a heavy cost on the other two goals of privacy and accuracy. The performance results have been compared with those of the unoptimized version of MASK on distorted database as well as with the well known algorithm Apriori for association rule mining on original database.