A secure privacy preservation model for vertically partitioned distributed data

Nowadays, data mining techniques are massively used via organizations intended for converting huge amount of data into information. Due to the advancement in database technology data are present at distributed sites consequently for carrying data mining analyzing in cost effective way we need to integrate these distributed data at one site. The predicament of anxiety here is that privacy of individual is at risk and we can find out sensitive information upon integration so a secure data model is needed to accomplish this task. In this paper, a new data model is proposed which integrates vertically partitioned distributed data securely by applying encryption on desire attributes, that cause privacy leakage and data owner, can also recover their data subsequent to integration. Further we apply CART algorithm on integrated data set and compared the result with C.45 algorithm to compare various parameter and shows that proposed algorithm give better performance than C.45.