High Impact Factor : 4.396 icon | Submit Manuscript Online icon |

Secure Mining of Association Rules in Horizontally Distributed Databases

Author(s):

Pratik Mendre , DR. D Y PATIL SCHOOL OF ENGINEERING, LOHEGAON; Sagar Pokharkar, DR. D Y PATIL SCHOOL OF ENGINEERING, LOHEGAON; Sukhada Vavhal, DR. D Y PATIL SCHOOL OF ENGINEERING, LOHEGAON; Shivanand Patil, DR. D Y PATIL SCHOOL OF ENGINEERING, LOHEGAON

Keywords:

Enhanced Privacy Data Mining, Distributed Computations, Frequent Item Sets (Caching), Association Rules

Abstract

We suggest a protocol for secure mining of association rules in horizontally distributed databases. The existing primary protocol is that of Kantarcioglu and Clifton [1]. Our protocol, like theirs, is rely on the Fast Distributed Mining (FDM) algorithm of Cheungetal, which is not a secured distributed version of the Apriori algorithm. The major ingredients in our protocol are two novel safe multi-party algorithms—one that calculates the combination of private subsets that each of the interacting players have, and another that tests the insertion of an element contained by one player in a subset contained by another. Our protocol offers enhanced privacy with respect to the protocol in [1]. In count, it is simpler and is significantly more effective in terms of interaction rounds, communication charge and computational cost. Data mining techniques are used to discover patterns in huge databases of information. But sometimes these patterns can disclose susceptible information about the data holder or persons whose information are the subject of the patterns. The idea of privacy-preserving data mining is to recognize and prohibit such revelations as evident in the kinds of patterns learned using traditional data mining techniques.[5].

Other Details

Paper ID: IJSRDV4I40094
Published in: Volume : 4, Issue : 4
Publication Date: 01/07/2016
Page(s): 1-4

Article Preview

Download Article