An efficient fuzzy keyword matching technique for searching through encrypted cloud data

Abstract

Cloud user intends to encrypt data before outsourcing sensitive data to the cloud which prevents data searching utility. Hence the necessity for searching through encrypted data appears. But in practice, it is very common for the user to misspell keywords while typing the words. Thus, fuzzy keyword search on encrypted data becomes an essential feature in searchable encryption. However, existing schemes suffer either from efficient handling of multi-letter errors or cannot distinguish anagrams. In this paper, we propose a scheme for fuzzy keyword search on encrypted data focusing on fuzzy word matching among dictionary words. Our proposed scheme construct a transformed keyword (monogram set) based on each letter and its position in the word, which enables to find out original word from its typo with maximum similarity matric. As a similarity metric, we have chosen a modified version of Jaccard similarity which ensures maximum similarity for the nearest word possibly the original one. Our experiment also suggests the applicability of our scheme.