论文部分内容阅读
【目的】通过对社交网站平台用户行为的分析,发现社会化小众群体中的核心用户,为社会化资源推荐服务提供参考。【方法】收集豆瓣读书用户的1 208个标签,对排名前100位的标签建立标签共现矩阵,分析用户的K-核网络结构,研究用户的K-核塌缩序列的波动情况。【结果】与度数中心度、最小K-核深度值等方法相比,基于K-核塌缩序列方法发现了新的社会化小众群体中的核心用户。【局限】样本数据规模较小且局限于某领域,排序问题不能得到很好的解决,需要进一步改进K-核分析方法。【结论】本研究有利于社交网站平台的管理者制定或改进新的资源推荐策略,从而促进社交网站平台更好地发展。
【Objective】 Through the analysis of the user behavior of the social networking platform, we found that the core users in the social minority groups provide a reference for the socialized resources recommendation service. 【Method】 A total of 1 208 labels of Douban readers were collected, and a label co-occurrence matrix was set up for the top 100 labels. The K-core network structure of users was analyzed to study the fluctuation of users’ K-nucleus collapsing sequences. 【Result】 Compared with the methods such as the degree center and minimum K-core depth, the core users in the new socialized minority groups were found based on K-core collapse sequence method. [Limitations] Sample data is small and confined to a certain area. Sorting problems can not be well solved, and K-nuclear analysis needs to be further improved. 【Conclusion】 This study is conducive to the social platform managers to develop or improve new resource recommendation strategies to promote the development of social platform.