ARTICLE | doi:10.20944/preprints202210.0064.v1
Subject: Mathematics & Computer Science, Information Technology & Data Management Keywords: Homomorphic Encryption; Privacy-preserving Record Linkage; approximate string matching
Online: 6 October 2022 (10:31:25 CEST)
String matching is an important part in many real world applications. It must robust against variations in string field. In record linkage for two different datasets matching should detect two patients in common in spite of small variations. But it becomes difficult in case of confidential data because sometimes data sharing between organizations become restricted for privacy purposes. Several techniques have been proposed on privacy-preserving approximate string matching such as Secure Hash Encoding etc. Relative to other techniques for approximate string matching Homomorphic encryption is very new. In this paper we have proposed a Homomorphic Encryption based approximate string matching technique for matching multiple attributes. There is no solution currently available for multiple attributes matching using Homomorphic encryption. We have proposed two different methods for multiple attributes matching. Compare to other existing approaches our proposed method offers security guarantees and greater matching accuracy.