Research On The Construction Method Of Expert Yellow Pages In Virtual Q&A Community

The rapid increase in the number of community-based question and answer services attracts lot of users.Most users are not only knowledge seekers but also knowledge providers.They reply to questions to share their knowledge.These experts are the carrier of implicate knowledge.Huge number of experts decrease the finding of required experts to ask for help directly or route questions for answers.In order to facilitate finding the suitable expert and alleviate information overload in finding process,in the paper,the expert yellow pages for CQA is constructed.First,considering the various length of texts,BTM model is used to model questions and experts.Secondly,the two-dimension expert yellow pages(2DEYP)are constructed.The dimensions are questions dimension and the expertise dimension.The intersections are the cluster of experts.The 2DEYP is expanded both laterally and vertically.Since the closer neurons represent similar topics,a novel labeling method is proposed for navigation.The method uses the distance between neurons as the discriminating capability.In order to alleviate the information overload further and discriminate the experts further,the ranking mechanism is proposed within each cluster.The sorting mechanism considers the professionalism of experts and the activity of experts.The professionalism of experts also includes professional depth and professional breadth.Experts' professional depth is calculated according to the number of praises between experts.User-user relationship network Figure,and then based on the modified Page Rank algorithm to get the expert's professional depth score;the expert's professional breadth is based on the JS distance algorithm,based on the knowledge of the field,the more other knowledge,the more Experts' breadth of expertise is stronger;expert activity considers the number and timing of answers.Finally,through the integration of expert professional depth,professional breadth and activity,the final ranking of experts is obtained through real data.The clustering method,the keywords extraction method and the expert ranking method are evaluated.The experimental results show that the method has good feasibility and good performance.