Privacy preserving association rules mining in vertically partitioned data
Kongling Wu · Jisuanji gongcheng yu sheji · 2012
To overcome the insecurity and the inefficiency of privacy preserving association rules mining in vertically partitioned data,a new privacy preserving association rules mining algorithm is presented.A new vector dot protocol is utilized by introducing inverse matrix and random numbers to hide sensitive information.To reduce the computational cost and improve the speed to generate the frequent itemsets,mining maximal frequent itemsets is combined with depth-first traverse strategy,and various pruning methods are also employed.The experimental results indicate the algorithm had better efficiency.