PRESERVING PRIVACY FOR PAGERANK ALGORITHM

Data mining has been emergingly applied in many fields to discover the knowledge from the huge data. To do that, information has been sent forward and backward among data owners, users, and maybe third parties. In this situation, it is necessary to design systems to exchange the data between the data owner, the client and third parties during data mining process without scarifying the sensitiveness of data. Hence, we need a privacy preserving mechanism while mining to protect the data as in the situation of sophisticated cyber-attack. In this work, we describe a model for ensuring the privacy in ranking on the graph using PageRank and Shamir Secure Sharing scheme. Specifically, Shamir Secure Sharing scheme has been applied to share the information of graph from the data owner to many servers (i.e. third party). Then, the share of graph on each server will be ranked separately. When the users need the results of ranking and make a request, the information from servers will be combined for the users. Doing this way, the third party doesn’t know the meaning of data but still run analyzing the data. Hence, data owner preserves the privacy of his data while users still retrieve a piece of the information as needed.