Privacy Preserving Data Mining

With the development of network, data collection and storage technology, the use and sharing of large amounts of data has become possible. Once the data and information accumulated , it will become the wealth of information. Data mining, otherwise known as knowledge discovery, can extracted “meaningful information” or “knowledge” from the large amounts of data, so supports people’s decision-making (Han & Kamber, 2006). However, traditional data mining techniques and algorithms directedly opreated on the original data set, which will cause the leakage of privacy data. At the same time, large amounts of data implicates the sensitive knowledge that their disclosure can not be ingored to the competitiveness of enterprise. These problems challenge the traditional data mining, so privacy-preserving data mining (PPDM) has become one of the newest trends in privacy and security and data mining research. In privacy-preserving data mining (PPDM), data mining algorithms are analyzed for the side-effects they incur in data privacy, and the main objective in privacy preserving data mining is to develop algorithms for modifying the original data in some way, so that the private data and private knowledge remain private even after the mining process (Verykios et al., 2004a). A number of techniques such as Trust Third Party, Data perturbation technique, Secure Multiparty Computation and game theoretic approach, have been suggested in recent years in order to perform privacy preserving data mining. However, most of these privacy preserving data mining algorithms such as the Secure Multiparty Computation technique, were based on the assumption of a semi-honest environment, where the participating parties always follow the protocol and never try to collude. As mentioned in previous works on privacy-preserving distributed mining (Lindell & Pinkas, 2002), it is rational for distributed data mining that the participants are assumed to be semi-honest, but the collusion of parties for gain additional benefits can not be avoided. So there has been a tendency for privacy preserving data mining to devise the collusion resistant protocols or algorithms, recent research have addressed this issue, and protocols or algorithms based on penalty function mechanism, the Secret Sharing Technique, and the Homomorphic Threshold Cryptography are given (Kargupta et al.,2007 ; Jiang et al., 2008; Emekci et al., 2007). This chapter is organized as follows. In Section 2, we introduce the related concepts of the PPDM problem. In Section 3, we describe some techniques for privacy preserving data mining In Section 4, we discuss the collussion behaviors in Privacy Preserving Data Mining. Finally, Section 5 presents our conclusions.