Total privacy preservation and search quality improvement in personalized web search
Leema Rose Mathew, Arun Elias, Chinnu Ravi · Journal of Web Engineering · 2016
Personalized Web Search (PWS) has dramatically improved the quality and effectiveness of web search nowadays. Its effectiveness totally depends on the user personnel data collection. But collecting personnel data possess a major threat of privacy risk. Even the search engine side can misuse this personnel information. Evidences show that users are more concerned about their privacy than search quality. We here propose a new system which offers improved search quality and complete privacy protection. A pattern based user interested topic presentation enhances the personalized search. When the user submits a query, our system creates a vector space model with a weight of query related frequent pattern words and then sent to the server side. To avoid the vulnerability of eavesdropping of this vector space model, a modified version of fully homomorphic encryption scheme is used. Experimental results of performance analysis show that our system improves the effectiveness of PWS and execution time.