Privacy preserving association rule mining based on homomorphic computations

Abstract

Privacy-preserving data mining (PPDM) techniques allow the knowledge extraction from data, while privacy preserving in data mining applications. Many of the researchers have recently made an effort to preserve privacy of sensitive knowledge or information in a real database. Association rule mining and frequent item-set mining are two popular and widely studied data analysis techniques for a range of applications. To ensure data privacy, in this paper, we design an efficient homomorphic encryption-based scheme for privacy preserving data mining. The main issues with some of the known privacy preserving methods are - high computational complexity and large communication cost required for their execution. Our methods provide perfect secrecy and resist various attacks to some extent in association rule mining process. We presented correctness, security analysis and experimental results for the proposed system. We also presented the comparison of our proposed method with other significant state of the art methods.