A toolbox for privacy preserving distributed data mining

Abstract

Distributed structure of individual data makes it necessary for data holders to performcollaborative analysis over the collective database for better data mining results.However each site has to ensure the privacy of its individual data, which means noinformation is revealed about individual values. Privacy preserving distributed datamining is utilized for that purpose. In this study, we try to draw more attention tothe topic of privacy preserving data mining by showing a model which is realistic fordata mining, and allows for very efficient protocols. We give two protocols which areuseful tools in data mining: a protocol for Yao?s millionaires problem, and a protocolfor numerical distance. Our solution to Yao?s millionaires problem is of independentinterest since it gives a solution which improves on known protocols with respect toboth computation complexity and communication overhead. This protocol can be usedfor different purposes in privacy preserving data mining algorithms such as comparisonand equality test of data records. Our numerical distance protocol is also applicableto variety of algorithms. In this study we applied our numerical distance protocol in aprivacy preserving distributed clustering protocol for horizontally partitioned data. Weshow application of our protocol over different attribute types such as interval-scaled,binary, nominal, ordinal, ratio-scaled, and alphanumeric. We present proof of securityof our protocol, and explain communication, and computation complexity analysis indetail.i