Hiding sensitive fuzzy association rules using Weighted Item Grouping and Rank based Correlated Rule Hiding Algorithm

Extracting knowledge from large amount of data while preserving the sensitive information is an important issue in data mining. Providing security to sensitive data against unauthorized access has been a long term goal for the database security research community. Almost all the research in privacy preservation is limited to binary dataset. Business and scientific data contain both quantitative and categorical attributes. The technique used for privacy preservation must ensure security of the database while maintaining the utility and certainty of the mined rules at highest level. This paper presents two techniques to hide quantitative sensitive fuzzy association rules - Weighted Item Grouping Algorithm and Rank based Correlated Rule Hiding Algorithm. Then the performance of the two techniques is evaluated based on the number of lost rules and ghost rules generated and how effectively the sensitive rules are hidden. The experimental results shows that the Rank based correlated rule hiding provides better performance than weighted item grouping in terms of side effects and number of modifications.