摘要
为了准确获取用户的需求,提出了一种基于移动数据的用户间影响力度量方法.首先,根据移动用户的交互行为构建社会网络,利用网络的拓扑结构和移动用户行为计算用户自身影响力.然后,根据用户自身影响力、用户间的交互行为和用户偏好相似度计算用户间影响力.计算基于用户交互行为的用户间影响力时,考虑了相应的上下文信息;计算基于用户偏好相似度的用户间影响力时,考虑了上下文信息以及用户偏好发生的先后顺序.最后,通过真实数据集与现有方法相比,得出本方法获取的用户间影响力的准确率更高.
In order to accurately obtain the user requirement, a calculating method of influence be- tween users based on the mobile data was proposed. First, according to the exchange behavior of mo- bile user, the mobile social network was constructed, and the influence of user oneself was calculated according to the topology of the constructed mobile social network and mobile user behavior. Then, the influence between users was calculated according to the user oneself influence, exchange behavior and the similarity of user preference between users. When calculating the influence based on the ex- change behavior, the context information was considered. When calculating the influence based on the similarity of user preference, the context information and the order of user preference occurred were considered. Finally, the experiment was executed using the real data set. The results show the pro- posed method can obtain more accurate influence between users than the existing methods.
作者
史艳翠
杨巨成
陈亚瑞
王建
Shi Yancui Yang Jucheng Chen Yarui Wang Jian(College of Computer Science and Information Engineering, Tianjin University of Science and Technology, Tianjin 300222, China Three Gorges Asset Management Center, Beijing 100069, China)
出处
《华中科技大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2017年第7期110-114,共5页
Journal of Huazhong University of Science and Technology(Natural Science Edition)
基金
国家自然科学基金资助项目(61402331
61402332
61502338
61272509)