Multi-keyword ranked search scheme for privacy protection in social networks

Abstract

In a social network friend-matching scenario, a proposed blockchain-based privacy protection scheme with cloud-edge collaboration aims to improve search efficiency while protecting the personal privacy information of dating users. The scheme encrypts and stores massive dating data on the cloud server (CS), and uploading the corresponding keyword index to the nearby edge node (EN) for distributed storage. To enhance the search experience for users, this paper uses the Bisecting K-means clustering algorithm to classify all dating documents, categorises the keywords with high relevance to form keyword grouping first, and then constructs the inverted index based on the keyword grouping results, so that we can quickly locate the position of the query keywords, and then find the matching document identifiers to speed up the ciphertext search. Simultaneously, this paper introduces the BM25 model to calculate the relevance scores of ciphertext documents to achieve the effect of ranking ciphertext documents. Under the general bilinear group model, the proposed scheme proves selectively secure against selective keyword attacks and plain text attacks. The simulation results indicate that this proposed scheme is a more efficient multi-keyword ranking search scheme for dating data.