ELPS:一种高效的微博信息传播轨迹提取算法

发布时间：2018-06-05 14:44

本文选题：社会网络 + 图挖掘　；参考：《计算机科学》2014年04期

【摘要】：近年来,随着社会性网络服务应用(SNS)的流行与发展,SNS已成为人与人之间重要的交流渠道。SNS中大量用户产生的数据内容包含了社会网络中信息传播的客观知识,由此SNS可用于研究社会网络中公众舆论的变化趋势及信息传播的相关规律。由于SNS服务中节点规模大、其用户间的信息传播通常出现离散而稀疏的情况,需要高效的信息传播观察手段。为解决该问题,提出信息传播轨迹用于研究社会网络中信息传播的基本规律,具体的方法为:(1)提出信息传播轨迹(info-trajectory)模型以记录社会网络中信息传播的具体路径;(2)针对微博社会网络,提出几个高效的信息传播轨迹抽取算法;(3)根据已获取的信息传播轨迹研究用户间转发信息行为的时序规律;(4)提出算法K-advocators-discover用于发现社会网络中促进信息传播的top-k名用户;(5)提供充分的实验测试来将所提方法用于抽取新浪微博上热点话题信息的传播轨迹,并采用K-advocators-discover算法分析新浪微博中促进信息传播的用户。实验结果验证,所提方法能高效地提取微博中信息传播轨迹,挖掘其中促进信息传播的用户。
[Abstract]:In recent years, with the popularity and development of social network service (SNS), SNS has become an important channel of communication between people. A large number of user-generated data content in SNS contains objective knowledge of information dissemination in social networks. Therefore, SNS can be used to study the changing trend of public opinion and the relevant rules of information dissemination in social network. Due to the large scale of nodes in SNS services, the spread of information between users is usually discrete and sparse, which requires efficient observation of information dissemination. In order to solve this problem, the paper puts forward the basic law of information transmission in social network. The specific method is: 1) to put forward the information transmission trajectory info-trajectorymodel to record the specific path of information transmission in the social network. (2) aiming at the Weibo social network, This paper puts forward several efficient algorithms for extracting information transmission trajectory. (4) A new algorithm K-advocators-discover is proposed to discover the top-k name of information transmission in social network. User No. 5) provides sufficient experimental testing to extract the trajectory of hot topic information on Sina Weibo by using the proposed method. K-advocators-discover algorithm is used to analyze the users who promote information dissemination in Sina Weibo. The experimental results show that the proposed method can efficiently extract the track of information propagation in Weibo and mine the users who promote the spread of information.
【作者单位】：中央财经大学信息学院计算机系;北京大学信息科学技术学院;
【基金】：国家自然科学基金(60970143,61272398) 北京市教委共建项目中央财经大学研究生教育改革项目资助
【分类号】：TP393.092

【参考文献】