Research On Privacy Protection Algorithm Based On Association Rules

As the continuous development of information technology, data showing explosive growth, it encourages data mining technology has been developed rapidly. However, with the application of data mining techniques data privacy problems ensued. With people’s attention degree of privacy increasingly high, privacy protection technology also gradually become a hot research. Privacy preserving data mining technology is a sensitive data protection application method in data mining. The main purpose of this technology is to ensure the accuracy of the data mining results while improving to sensitive data, knowledge of privacy protection. It is also now the main direction of the people in the study of privacy protection.This paper mainly studies on association rules mining algorithm privacy protection. The paper introduces the association rule mining and privacy-related concepts and technologies. And it analyzes the existing two kinds of Privacy Protection algorithms based on association rules. For the lack of time efficiency and rules loss rate on the algorithms, we propose improved methods:(1) In this paper, we make a detailed analysis based on Randomized Response with Partial Hiding (RRPH) method. For the inadequate of the efficiency of support reconstruction on this method, we use divide and conquer strategy to optimize the solving of the transformation matrix in support the reconstruction. And we use the set operations methods to optimize the counting of the distorting data set support. Finally, according to experiments and analysis, demonstrate the effectiveness of the improved algorithm.(2) In this paper, we analysis the privacy protection of association rule mining (PPARM) algorithm. Found that the loss rate of rules is relatively high, while the privacy protection method hiding sensitive rules. To solve this problem, this paper proposes a privacy protection algorithm based on heuristic (HPP). The algorithm improves the method of selecting victim items in the transactions of the data sets. It ensures minimal impact on non-sensitive rules, while hiding sensitive rules. Finally, the experiment proved HPP algorithm reduces the loss rate rules. It showed the effectiveness of the algorithm.