The study of information dissemination and prediction in social media data is a hot spot of public opinion analy-sis.In this paper,we take the Twitter data as an empirical research to find the factors of predicting information propaga-tion.We design algorithm to analyze the influence of active neighbor nodes on the forwarding behavior,and propose three structure of active neighbor nodes to verify the influence on the forwarding behavior.We use start random walk algorithm to study the influence on the forwarding behavior between the reply and the mention.Finally,we implement the proposed method in twitter data sets.The social behaviors such as following,replying,mentioning are taken into consideration.%基于社会化媒体数据研究信息的传播及预测，是当前网络舆情分析的一大热点。以 Twitter 数据为研究对象，以探寻影响信息转发的因素为研究目的，设计算法分析活跃邻居节点数对转发行为的影响，提出3种活跃邻居节点结构并验证三者对转发行为的影响程度，运用重启动的随机游走算法研究回复与提及关系对转发行为的影响。在Twitter数据集上实现了该算法，证明了活跃邻居节点数、活跃邻居节点结构、回复与提及关系对转发行为的影响。
展开▼
