科研社交网络中基于链接预测的专家推荐研究

Expert Recommendation in Scientific Social Network Based on Link Prediction

随着Web2.0理念和技术的引入,科研社交网络为科研人员之间进行交流和协作提供了一个新的便捷平台。但近年来随着科研社交网络的不断发展,如何在海量专家信息中找到自己感兴趣的专家并与之建立合作关系变得十分困难。为此,通过综合分析科研社交网络中专家所具有的知识信息以及社会关系信息,并以此为基础,构建链接预测模型对科研社交网络中的用户进行相关专家推荐。最后,选取科研社交网络Scholar Mate平台进行实验,验证了本文提出方法的有效性。

With the introduction of Web 2.0 concept and technology, scientific social network provides a new and convenient platform for researchers to communicate and collaborate with each other. However, with the development of scientific social network, it becomes more and more difficult for researches to find their interested experts from the huge amount of expert information. Therefore, this paper tried to analyze experts＇ knowledge information and social information in the scientific social network,and then a link prediction model for expert recommendation in the scientific social network was proposed. Lastly, ScholarMate, one of the popular scientific social networks, was select- ed to conduct the experiment. Experimental results verified the effectiveness of the proposed method.