Hybrid optimization techniques for privacy preserving in data mining

Abstract

newlineCorporations are increasingly maintaining large size electronic databases which are accessed using internet or intranet Important information utilized from the data using data mining methods While performing data mining steps there is an inherent danger to the privacy of the data The sensitive data available in the database should not be accessible to all users Techniques proposed in the literature for privacy protection Privacy Preserving Data Mining PPDM use algorithms on confidential data that are unknown even to the algorithm operator Users personal information newlineand information concerning their collective activity are two major privacy preservation dimensions The former is termed individual privacy preservation and the latter collective privacy preservation Most of the privacy preserving methods are based on reduction in the granularity of the representation of the data This leads to information loss but improves privacy Therefore in PPDM there is a trade off between loss of information and the privacy Effective techniques are required which do not compromise the security mechanisms Some of the techniques proposed for privacy preserving include randomization method k anonymity model newlinel diversity and distributed privacy preservation newline