Social Network Inference of Smartphone Users Based on Information Diffusion Models
Social Network Inference of Smartphone Users Based on Information Diffusion Models
复制标题
DOI：
10.1007/978-3-642-25856-5_23
复制
发表时间：
2011-12
期刊：
影响因子：
--
通讯作者：
Tomonobu Ozaki;M. Etoh
中科院分区：
文献类型：
--
作者：
Tomonobu Ozaki;M. Etoh
在本文中，我们提出了推断智能手机用户社交网络的模型。通过将信息扩散模型的概念应用于智能手机中应用程序执行的日志，用户之间关系的强度将被估计为一个优化问题。利用时差函数和应用重要性函数来准确捕捉用户行为。此外，用户的隶属关系信息作为外生因素被有效利用。使用157个智能手机用户的实验结果表明，所提出的模型优于朴素方法，并适当地推断出社交网络。特别是，该模型成功地准确捕获了用户群体中的重要关系。
In this paper we propose models for inferring a social network of smartphone users. By applying the concept of information diffusion models to the log of application executions in smartphones, strength of relationships among users will be estimated as an optimization problem. Functions on time difference and application significance are employed to capture user behavior precisely. In addition, affiliation information of users is effectively utilized as an exogenous factor. Experimental results using 157 of smartphone users indicate that the proposed model outperforms naive methods and infers a social network appropriately. Especially, the model succeeds in capturing the important relations in user communities accurately.