Secure Semantic Search Based on Two-Level Index Over Encrypted Cloud

With the rapid development of the IoT and the mobile applications, users are tending to outsource the data to the cloud servers. Thus, the encrypted data search on the cloud is very important. Now there are a lot existed keyword search schemes in which the relationship between the size of the file set and the search time is linear. In order to solve this problem, we propose a semantic retrieval framework for central word expansion based on two-level index. In this paper, we have taken a new approach for index construction - the two-level index to ensure that the retrieval time is not affected by the file size. In order to better meet the semantic requirements of user queries, we introduced the central word expansion technology to further improve the accuracy of the search. The main idea of central keyword extension semantic search (CKESS) based on two-level index is that match the expanded central keyword with index firstly, then compute the similarity between the query and the index under the first matching result, and finally return the result with the highest similarity. Our proposed solution meets the privacy protection requirements under two different threat models. Through the experiment of the real data set, we prove that our scheme is efficient, accurate and secure.