Private Record Matching for Unstructured Data Using Hybrid Approach

Volume 11 Issue 4 April - June 2017

Research Paper

Private Record Matching for Unstructured Data Using Hybrid Approach

S. Rama Krishnan*

Department of Information Technology, Kongu Engineering College, Erode, Tamil Nadu, India.

Krishnan, R. S. (2017). Private Record Matching for Unstructured Data Using Hybrid Approach. i-manager’s Journal on Software Engineering, 11(4), 24-29. https://doi.org/10.26634/jse.11.4.13818

Abstract

Record Matching is the task of identifying records which are present across different databases. Some approximate techniques should be available to match the records if unique identifiers are not present. The record matching consists of mainly two approaches, such as sanitization approach and cryptographic approach. The sanitization technique mainly relies on methods, such as K-anonymization and random noise addition. The cryptographic technique uses Secure Multiparty Computation (SMC). It will provide accurate results, but the cost involved will be very high. The hybrid technique combines both the sanitization and cryptographic approaches. The major advantages of combining these approaches are it uses effective decision rule that matches the input values. Based on this, the privacy provided will be high. But the cost involved will be very high and is not effective. Hence the proposed method uses the two party protocols which involve record matching only between the data parties. There will be no trusted third party involved and the secure matching is done by using Binning method.