一种在线社会网络的信息扩散预测模型

A Prediction Model of Information Diffusion in Online Social Networks

论文针对在线社会网络中的信息扩散问题,提出了一个信息扩散的预测模型。首先给出了以好友关系作为用户距离度量的方法;然后将信息扩散看作是"社会扩散"和"内部扩散"两种方式同时作用的结果,并分别将Fick扩散理论和Logistic增长模型用于描述这两个过程,设计了Fick-Logistic扩散预测模型。最后,用该模型对Digg数据集中最具代表性的新闻实例进行预测。较高的预测准确率表明,论文提出的Fick-Logistic扩散预测模型能较好描述Digg在线社会网络中的信息扩散过程,具有较好的预测性能。

This paper proposes a diffusive model named Fick-Logistic model to describe and predict information diffu- sion in online social networks. Firstly, paper gives out the definition of distance between a pair of users in network graph by using shortest path measured by friendship hops. Then paper considers the information diffusion in online social networks as two separated process, social process and growth process. And it applies Fiek diffusion theory and Logistic growth model in- to information diffusion process in the context of online social network. Lastly, the paper utilizes proposed model to predict diffusion process of the most representative news story in Digg dataset. The high prediction accuracies indicate that the model can precisely describe the process of diffusion in Digg and has good capability of prediction.