识别网络论坛中有影响力用户

发布时间：2018-03-13 02:36

本文选题：网络论坛　切入点：影响力　出处：《计算机研究与发展》2013年10期 　论文类型：期刊论文

【摘要】：网络论坛已经成为网络用户发布信息的重要渠道.在论坛中对热点话题的讨论影响着物理世界中人们的看法、观点以及国家政策法规的制定.由此提出一系列研究问题:如何计算用户影响力?不同主题不同时间下用户影响力如何比较?用户影响力发展趋势如何?根据幂律规律,大量用户形成"长尾",如何识别有影响力用户?以主题为单位,提取用户间回复关系,构建用户对话关联图,回复次数和回复长度形成用户行为特征,入度和出度形成网络结构特征.在Pagerank算法基础上,结合用户行为特征以及用户间关联网络特征,提出基于多属性的用户影响力排序算法(multiple attributes rank,MAR).并依据发表时间进行时间段切分,得到论坛上每日有影响力用户排行榜,进一步分析了有影响力用户演化趋势.以天涯网络论坛真实数据进行实验,从多角度评价有影响力用户以及MAR排序算法,得到一些有趣结论并对未来工作进行了展望.
[Abstract]:Internet forums have become an important channel for web users to publish information. The discussion of hot topics in the forums has influenced the views of people in the physical world. From this point of view and the formulation of national policies and regulations, a series of research questions are raised: how to calculate the influence of users? How do you compare the influence of users under different themes and times? What is the trend of user influence? According to the law of power law, a large number of users form a "long tail", how to identify influential users? Based on the topic, the response relationship between users is extracted, and the user dialog correlation graph is constructed. The response times and the reply length form the user behavior characteristics, and the network structure features are formed by the degree of entry and the degree of output. Based on the Pagerank algorithm, the network structure is formed. Combined with user behavior characteristics and user association network features, this paper proposes a multiple attributes ranking algorithm based on multiple attributes, and divides the time period according to the time of publication to get the daily influential user list on the forum. In this paper, the evolution trend of influential users is further analyzed. Based on the real data of Tianya Network Forum, the influential users and MAR sorting algorithm are evaluated from various angles, and some interesting conclusions are obtained and the future work is prospected.
【作者单位】：哈尔滨工业大学计算机网络与信息安全技术研究中心;
【基金】：国家“八六三”高技术研究发展计划基金项目(2010AA012504) 国家“九七三”重点基础研究发展计划基金项目(G2011CB302605) 国家自然科学基金项目(61173145)
【分类号】：TP393.09

【参考文献】