基于PageRank的微博排名MapReduce算法研究

摘要

With the development of social network service,mining and analyzing data from SNS is becoming an active area of science. In micro blog,the user ranking is based on the number of fans,but it is not very fair. In this paper,propose a new ranking algorithm based on web PageRank,in which use the data from Sina Weibo to yield a graph with nodes and edges. Then build a transition probability ma-trix to compute every user’s PageRank. This algorithm can make the user ranking more fair and more closely to reflect the reality. The experiments are conducted in cloud,which present a new ranking result and the algorithm has some practical value,comparing with the follower ranking.%　　随着社交网络的发展，对于其数据的挖掘与分析已经成为一个热门领域。在微博中，用户排名通常是单纯根据粉丝人数进行排列，而这种方法并不公正。针对这一问题，结合网页PageRank算法，提出了新的排名算法，以用户为节点，用户关系为有向边，建立概率转移矩阵，计算微博用户PageRank值。该算法能有效减少垃圾用户对微博排名的影响，来提高排名的公平性与准确性。实验测试在云环境下进行，结果显示了新的排名结果，与现有的微博粉丝排名相比，更加公平，具有一定的实用价值。