Privacy Preserving Data Mining Using Secure Multiparty Computation Based On Apriori and Fp-Tree Structure of Fp-Growth Algorithm
P.Yoganandhini1, G.Prabakaran2
1P.Yoganandhini*, Research Scholar, Department of Computer and Information Science, Annamalai University.
2Dr.G.Prabakaran, Associate Professor, Department of Computer Science and Engineering, Annamalai University.
Manuscript received on December 16, 2019. | Revised Manuscript received on December 22, 2019. | Manuscript published on January 10, 2020. | PP: 353-359 | Volume-9 Issue-3, January 2020. | Retrieval Number: B6866129219/2020©BEIESP | DOI: 10.35940/ijitee.B6866.019320
Open Access | Ethics and Policies | Cite | Mendeley
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: In this work, a method is proposed to deal with secure multiparty computation (SMC) based problems. The computation is done on the grocery dataset collected from three various grocery shops. The privacy is maintained by generating the rules based on FP-Tree algorithm under Association Rule Mining (ARM). Privacy and correctness are the important requirements of SMC. In privacy requirement, the things apart from necessary are not learned. This implies that only output will be learned by the parties. Each party must receive correct output to ensure the correctness. In this work, secure auction is done using SMC and frequent item sets are computed to perform the association rule mining. The most familiar FP-growth schemes have the short fallings like former space complexity and latter time complexity. The performance of the algorithms has been enhanced by using APFT algorithm which is a combined version of FP-tree structure of FP-growth algorithm and Apriori algorithm. The conditional and sub conditional patterns are not generated continuously in APFT. The speed of the APFT is high when compared to Apriori algorithm and FP-growth. The correlated items are included by modifying APFT and non correlated item sets are shaped by using APFT. This modification is used for FP-tree optimization. From the frequent item set, the loosely associated items are removed by using this modification. The system implemented is clearly described and its performance is evaluated. The results confirmed that the proposed scheme is extremely effective.
Keywords: Apriori algorithm, Association Rule Mining, Distributed Data Mining, Frequent Pattern Growth, Privacy Preserving Data Mining (PPDM), Secure Multiparty Computation (SMC).
Scope of the Article: Data Mining