论文部分内容阅读
【目的】为全面获取专家资源,探究“小众专家”特征识别方法。【方法】以知名社交博客MetaFilter为例,利用用户社交活动数据,构建用户关系网络,统计节点网络结构指标:中介中心度和聚集系数。结合聚类分析和时序分析,判别不同时期节点特征及角色。【结果】综合类群网络特征,获取“小众专家”集合,依据集合时序变动情况细化“小众专家”分类。【局限】只对music版块评论关系进行角色判别及迁移分析,未来工作将扩展至更多版块,对比分析不同语义环境下“小众专家”类群“稳定–变化”特点。【结论】“小众专家”是对现有专家集合的有效补充,其识别研究可用于专家团队构建、专家推荐、专家检索等方面。
【Purpose】 In order to gain full access to expert resources, explore the “niche experts” feature recognition method. [Methods] Taking the well-known social blogging MetaFilter as an example, the user relationship network was constructed by using the data of social activities of users, and the network structure index of the node was calculated: intermediary center degree and aggregation coefficient. Combined with cluster analysis and time series analysis, the characteristics and roles of nodes in different periods are identified. 【Result】 The results show that the “niche experts” can be collected according to the characteristics of the network. Based on the changes of the aggregation timing, “niche experts” are classified. [Limitations] Only judge and relocate the relationship between the commentaries on the music section, the future work will be extended to more sections, and the characteristics of “minority experts ” “stable change ” will be compared and analyzed under different semantic environments. 【Conclusion】 “niche experts ” is an effective complement to the existing expert collection, and its identification research can be used for expert team building, expert recommendation, expert search and so on.