中文摘要：

研究了在线社会网络中的多源点信息扩散问题。首先针对在线社会网络中的多源信息扩散给出了以好友关系为度量的用户之间距离的定义;然后基于Digg数据集设计了一种多源信息实例的近似选取算法,并通过大量实验分析研究了多源信息扩散实例的扩散规律和特征;最后利用一种线性扩散模型对多源信息扩散进行了预测。较高的预测准确率表明,提出的距离度量方式和多源信息选取算法是可行有效的,并证明了该线性扩散模型对于多源信息的扩散具有较好的预测性能。

英文摘要：

This paper discussed the issue of information diffusion initiated from multiple sources in online social network.Firstly,it gave out the definition of distance between a pair of users in network graph by using shortest path measured by friendship hops. Then it designed an algorithm for selecting multiple sources news stories in Digg dataset,and conducted a series of empirical study to reveal the intrinsic features and spatial-temporal characteristics of multi-source information diffusion in online social network. Lastly,it utilized a mathematical model named linear diffusive model to predict multi-source information diffusion process in Digg. The high prediction accuracies indicate that this distance definition and multiple sources selecting method are both feasible and effective,and further validates the performance of predicting of the linear diffusive model.