Social Question Answering Community Respondent Discovery Research

With the in-depth popularization of the Internet among the people,the continuous increase in the number of Internet users,and the further development of information technology,the social Q&A community has become a popular place for people to quickly search and exchange information and knowledge.It can break through the boundaries of time and space and meet knowledge acquisition needs and communication needs in fragmented time of users,which has adapted to the fast pace of work and life of modern people.However,in the social Q&A communities with voluntary participation,such as foreign Yahoo! Answers and Stack Overflow,domestic communities like Zhihu Community,Sogou Ask,Baidu Know,etc.,there are still problems that users have asked for and have not been answered for a long time,Or the problem that the question is not answered with professionalism,completeness,and satisfaction.Over time,the questioner may have frustration,reduce expectations of the Q&A community,and even leave the platform.This is obviously not conducive to the sustained and healthy development of the Q&A community.Therefore,how to retain users from the level of knowledge sharing in the Q&A community and seek expert users with high possibility of answering questions,in order that new questions in the community can be effectively and professionally answered is the focus of this paper.Finding professional answerer with a high probability of answering questions can meet the expectation of the questioner to get fast,professional and convincing answers,shorten the waiting time for questioners,and help the healthy and sustainable development of the social Q&A community,which has practical significance to a certain extent.Previously,some scholars have discussed how to identify expert users in a topic area of the Q&A community,or study the influencing factors of user knowledge sharing in the community,or improve related algorithms.In fact,if expert users are restricted by various conditions,they cannot When answering questions in a timely manner,it's still difficult to change the status in the community where questions are not answered orsatisfactory responses are not received.Therefore,this article uses motivation theory and social capital theory to combine the relevant research found by experts,and strives to find professional answerers with greater possibility of answering the questions to solve the above problems.In addition,this paper has enriched the related academic research to a certain extent and can also provide some reference for the research of expert recommendation,problem routing and other topics,so this paper has certain theoretical and practical significance.This study sets up the dataset required for the experiment based on the personal information and related question and answer information which captured by the web crawler under the medical topic of Zhihu community,and analyzes the possibility of users answering questions in the community with social capital theory and motivation theory.The research model is constructed to solve the following three questions:(1)How to identify professional users on a specific topic in the online Q&A community?(2)What kind of users are more likely to answer questions?(3)How can I find users who are more likely to answer questions among the identified expert ? This research uses the machine learning modeling.Based on the user's background information,the user's interactive information behavior in the Q&A community,and the user's activity index in the Q&A community,the user's credibility is evaluated by means of general mathematical methods and TOPSIS method;use tf-idf,LDA topic model and general mathematical methods to analyze users 'professional topics in their historical answers and articles in order to analyze the user's professionalism;based on the social relationship network formed by the users in the community,with the help of PeopleRank algorithm we analyze the user's importance in the community.We also combine with the power of the user's voice in the community so as to measure the user's authority.In the study,the relevant parameters of the model were debugged and optimized according to the experimental data,in order to obtain better experimental results.In addition,this paper also conducted a comparative experimental analysis with the classic algorithms PageRank and HITS.The experimental results verify the effectiveness and superiority of the research method and model in this paper.