Research On The Recognition Of User Interests Of Social Q & A Sites

The social Q & A Site is a social platform that provides the opportunity of mutual communication and help for users. After years of development,social Q & A Site has accumulated a lot of information and become an important channel for users to access information, it is convenient for users to acquire knowledge from the platform. But there are also some problems for users, such as "information overload", the low quality of answer, Q & A cycle is too long, single search function and so on. It is urgent for social Q & A Sites to improve service quality and customer satisfaction. Research on the recognition of User Interests of social Q & A Sites is necessary, which could provide personalized service based on user characteristics, improve the level of service and enhance user stickiness.This paper researched user interests in Zhihu, a social Q & A Site in China, it is organized as follows. The first chapter introduces the research background and significance on the recognition of User Interests in social Q & A Sites and analyzes current research status, then introduces the research contents and research methods in this paper. The second chapter summarizes the related theories of User Interest Recognition,including the connotation, characteristics and operating mechanism of social Q & A Sites,the connotation, characteristics and representation of User Interest, the related technologies of User Interest extraction, and then. The third chapter discusses the recognition process of User Interest based on existing research. In this chapter, the characteristics of User Interests in Zhihu are analyzed, which is divided into explicit interest and implicit interest. Then a user interest representation framework based on keyword according to its characteristics is constructed. Later the steps of extracting user's explicit interest from the tag of user's questions and the steps of extracting user's implicit interest from user's answers are introduced to get two sets of keywords and its weight.Finally, the weights of the two sets of keywords and extract keywords with higher weight are recalculated to represent the user interest, and user interest model is evaluated by the method of phrase similarity calculation.The forth chapter demonstrates the construction method of the User Interest Model through the experimental way. Which includes creating a data set by collecting user data from Zhihu, analyzing the data set and evaluating the experimental results. And this experiment verifies the feasibility, validity and accuracy of the method.