Multilevel Privacy Preserving in Distributed Environment using Cryptographic Technique

—Data mining extracts pattern or knowledge from a large amount of data. In the applications that are based on information sharing, an additional challenge is faced; while dealing with data containing sensitive or private information. Common data mining techniques do not address this problem. Therefore, the knowledge extracted from such data may disclose patterns with sensitive or private information. This may put the privacy of individuals or groups, the business strategies and classified information at risk [1]. In the recent past years, Privacy Preserving Data Mining (PPDM) has attracted research interest with potential for wide applications. This paper considers all data repositories and one node for pattern extraction at distributed locations. Many techniques like anonymity, randomization and cryptography have been experimented with privacy preserving data mining. This paper considers information system based approach as not all attributes may store same level of sensitive data. Therefore, some attribute values may require higher degree of privacy preservation than some others. This paper explores the use of cryptography, namely DES algorithm for encrypted data sharing to achieve privacy preservation. The experiments have been carried out on the size of the secret key that is proportional to the level of sensitivity of the attribute for multilevel security. The RSA algorithm has been used to encrypt the DES algorithm key for enhancing secured communication.