基于Hadoop云平台的新浪微博社交网络关键节点挖掘算法

Key nodes mining algorithm in Sina Weibo social network based on Hadoop cloud platform

为了高效地分析挖掘新浪微博社交网络信息传播过程中的关键节点,以Hadoop云计算系统作为存储和处理平台,在X-RIME大规模社会网络分析工具开源框架基础上,针对社交网络中使用HITS(hypertext induced topic selection)链接分析算法挖掘关键节点时,未能体现节点和连接的社会属性问题进行改进.新算法充分考虑了社交网络节点和边的社会属性,对HITS算法节点和边的社会属性权值进行优化计算,提出适合社交网络特点的加权HITS算法.通过Hadoop云平台分别运行加权HITS算法和传统HITS算法对新浪微博社交网络数据进行分析.实验结果表明,加权HITS算法比传统HITS算法具有更高的执行效率和结果区分度,加权HITS算法更适合于大规模社交网络信息传播过程中关键节点的分析挖掘.

To efficiently analyze and mine the key nodes in the information dissemination process of Sina Weibo social networks,the Hadoop cloud computing system is used as the storage and process platform,based on the X-RIME massive social network analysis open source framework,the traditional hyperlink analysis algorithm HITS(hypertext induced topic selection)is improved by exploring the social attributes of nodes and edges.Based on the social attribute characteristics of the nodes and edges in social networks,the social attribute weight values of nodes and edges are computed and optimized in the new weighted HITS algorithm.The weighted HITS algorithm and the traditional HITS algorithm were implemented to analyze the Sina Weibo dataset in the Hadoop cloud platform.Experimental results show that the weighted HITS algorithm provides higher efficiency and better discrimination than the traditional HITS algorithm,and the weighted HITS algorithm is more suitable for analyzing and mining the key nodes of the information dissemination process in large-scale social networks.