Research On The Encrypted Data Supporting Synonymous Multi-keyword Fuzzy Search In The Cloud Computing

Cloud storage is one of the most widely used applications in the cloud,which provides users with scalable and resilient storage services through the Internet.More and more people outsource critical and personal data to cloud servers,resulting in that maintaining the confidentiality and integrity of these data is crucial.We can encrypt data before outsourcing to protect data,but it becomes a problem that retrieving the required files from the encrypted cloud.Searchable encryption technology can search cipher text through keywords while ensuring the privacy of user data,aimed at locating related encrypted files through keywords efficiently and accurately for users.The proposed schemes have achieved better retrieval efficiency and accuracy for single keyword and exact keyword,but the implementation of multi-keyword fuzzy search is still not satisfactory.Current fuzzy search schemes rely on extending index to cover possible keyword misspellings,which will result in larger index files and higher search complexity.This paper proposes a scheme to support synonyms and multi-keyword fuzzy ranking retrieval in encrypted cloud data.This solution enhances the user's search experience by providing fuzzy and synonyms multi-keyword retrievals.We use the Natural Language Processing to expand the synonym of the keyword to convert the synonym search into an exact match,and we construct the index through Bloom Filter based on Local Sensitive Hash functions to achieve fuzzy multi-keyword search,at the same time,it can minimize the overhead of updating the index when dynamic deleting and adding a file.Finally,we sort the search results by inner product relevance and return the most relevant cipher file to the user.This paper uses hash algorithm and secure k-nearest neighbor algorithm to effectively resist the selection of keyword attacks,and achieves the retrieval of cipher text in the semi-trusted cloud server.The main work of this article is as follows:(1)This paper studies several existing schemes supporting multi-keyword fuzzy retrieval,and analyzes its index construction,keyword matching and cipher text sorting methods,and lists the advantages and disadvantages of each scheme.(2)This paper expands synonym set for keywords,and uses the Bloom Filter based on the local sensitive hash functions to build the index.Insert the weight of the keyword to match fuzzy words,which improves the retrieval efficiency and the accuracy of the sorting.(3)Using a secure encryption network backup system as a platform,this paper achieves searchable encryption in the cloud storage by using python language,and verifies the feasibility and security of the scheme.(4)Based on designing and implementing system supporting encrypted data retrieval,this paper analyzes the security and construction efficiency of the index structure and the trapdoor structure.The retrieval efficiency of the scheme is tested with the actual dataset,and the security and accuracy of the scheme are analyzed in detail.