In distributed data mining, secrecy of private data input of parties with similar background, is achieved by Secure Multi Party Computation (SMC). One of the mostly used tool of SMC is secure sum protocol which has been modified by researchers using many techniques to provide utmost security. In this paper, we propose another novel secure sum protocol to provide more data security in an efficient way named Double Random Partitioned Model (DRPM) protocol for multi-party computation that uses the collaboration of data segmentation, value randomization technique and trusted third party for ensuring zero data leakage among participating parties. Proposed method have reduced computational steps noticeably than all other existing protocols. The comparative study shows that the proposed protocol performs much better than the existing protocols in terms of communication complexity and computation complexity, e.g., proposed DRPM protocol improves 85% on computational complexity over the existing best one.

No abstract