Research And Implementation Of Privacy Protection Method For Personalized Search

With the rapid growth of Internet information,personalized search has become one of the important technologies to improve user search efficiency.The essence of personalized search is to build a user interest model through user history search records to provide users with accurate search services.However,the user privacy leakage problems caused by this not only harms the user's interests,but also makes the personalized search service encounter an unprecedented crisis of trust.Therefore,personalized search privacy protection technology emerges as the times require.However,there are still some shortcomings in the research of existing personalized search privacy protection.If there is no difference in content sensitivity,the same privacy protection method is adopted for each search;user query preference and cocurrent are the problems that cause the user's query intention to leak.In view of the above problems,this paper proposes a privacy protection method for personalized search.The main work is as follows:Firstly,in view of the fact that the existing protection methods do not distinguish the content sensitivity,and the user uses the same privacy protection method for each search,this paper proposes a privacy protection algorithm for sensitive search content.The user dynamic interest model is established by the hierarchical tree algorithm and the ODP(Open Directory Project).Then the keyword search is performed on the user search content based on the TextRank algorithm,and the extracted keywords are compared with the user interest model to determine the user search theme.Based on the extracted topics and user information sets,sensitivity constraints are established,and the sensitivity of the search content is calculated.The sensitivity of the search contents is differentiated.Different intensity privacy protection methods are adopted according to the sensitive value.Through theoretical and experimental analysis,the algorithm realizes the differentiation of the sensitivity of the search contents,and provides different intensity privacy protection for the user search based on the sensitive value.Secondly,this paper proposes a privacy protection algorithm that confuses the intent of the query,in view of the user's query preference and co-occurrence,which leads to the user's intention to disclose the problems.By introducing a user memory decay model to construct a dynamic confusion set,the sensitivity of the query record is kept dynamically updated,and a confusion set is generated using low-sensitive query records.In addition,for the co-current behavior existing in the user's continuous search for content,the co-occurrence words of the search content are extracted,and similar semantic substitution is performed.The random confusion algorithm based on the random number is used to confuse the user's query intent,so as to achieve effective protection of the user's query intent.Through theoretical analysis and experimental comparison,the algorithm achieves effective protection of user's query intent.Finally,the algorithm proposed in this paper is used to design and implement a privacy-protected personalized search system,which mainly includes sensitive content analysis,query confusion and privacy protection,and tests the system's functions and performance.The system effectively validates the personalized search privacy protection method proposed in this paper,realizes the sensitive differentiation of user search content,effectively confuses the user's query intention,and provides safe and reliable protection for the user's personalized search.