Study and design of privacy Preserving data mining algorithms

Abstract

The modern developments in the area of data collection storing and newlinenetworking have led to data stored and maintained in different locations Data newlinemining algorithms help extraction of hidden knowledge from large data newlinerepositories The data is not always present in the centralized server The data newlineis available in single or different locations The data is partitioned vertically newlineor horizontally or both and distributed among different locations. Sharing data newlinerepositories present in different locations during mining process creates newlineunease on privacy of data Privacy preserving data mining is a technique that newlineprotects the sensitive data from being revealed to others during a mining newlineprocess The most popular privacy preserving data mining techniques are newlinesimple multi party computation perturbation and anonymization Secure newlinemultiparty computation SMC allows the parties at different locations to newlinecarry out distributed data mining tasks without revealing any additional newlineprivate information SMC protocol mostly uses encryptions which provide a newlinegood level of security and increase in protocol complexity Perturbation newlinedistorts the original sensitive data before it is used for mining Distortion of newlinethe data can reduce the acc newline newline