An Efficient Algorithm for Privacy Preserving Maximal Frequent Itemsets Mining

This paper addressed the insecurity and the inefficiency of privacy preserving association rule mining in vertically partitioned data. We presented a privacy preserving maximal frequent itemsets mining algorithm in vertically partitioned data. The algorithm adopted a more secure vector dot protocol which used an inverse matrix to hide the original input vector, and without any site revealing privacy vector. The mining strategy was based on depth-first search for the maximal frequent itemsets. Performance analysis and experimental analysis showed that the algorithm possessed higher security and efficiency.