Абстрактный

Privacy Preserving CART Algorithm over Vertically Partitioned Data

Raghvendra Kumar, Ashish Jaiswal, Divyarth Rai

Data mining classification algorithms are centralized algorithm and works on centralized database. In this information age, organizations uses distributed database. Since data mining of private data is one of the keys to success for an organization, it is a challenging task to implement data mining in distributed database. Collaboration of different organization brings mutual benefits to the party involved. So different organizations wants to collaborate and execute efficient data mining algorithm. This arises privacy issues. Organizations are unable to collaborate because the privacy of private data is not fully preserved. In this paper CART algorithm is implemented over vertically partitioned data. For efficient privacy preservation of private data, privacy preserving protocols such as scalar product, x (ln x) protocols are used.