Abstract: Driven by new life experiences and social influence, the preferences adopted by individuals are con-stantly modified over time. In order to make up for the shortage that lots of existing geo-social networks (GSNs) models cannot capture dynamic preferences of user, this paper builds a temporal geo-social network model which can detect user dynamic preferences, and a dynamic preference value model to quantify user dynamic preferences. Accordingly, this paper proposes a dynamic preference-based group query (DPG). For optimizing the efficiency of DPG query algorithm, this paper also proposes UTC-tree to index user temporal check in records. UTC-tree can avoid traversing all users' check-in records to speed up the evaluation of user dynamic preferences. Finally, a well-interactive DPG query system employing the DPG query algorithm is implemented. A large number of comparison experiments prove the validity of UTC-tree and the scalability of DPG query algorithm.

Key words: temporal geo-social network, dynamic preference, query index, group query, time window

摘要： 在新生活体验与社交影响的驱动下，用户偏好随时间不断改变。为弥补现有的地理社交网络模型往往无法获取用户动态偏好的不足，构建了能够检测用户动态偏好的时序地理社交网络模型，以及用于用户动态偏好评估的动态偏好值模型，并提出了一种基于动态偏好的组查询（DPG）。为优化DPG查询算法效率，设计了UTC-tree索引用户时序签到记录。UTC-tree避免了在查询中遍历全部的用户签到记录，加速用户动态偏好评估。最后，采用DPG查询算法实现了交互良好的DPG查询系统，并通过大量对比实验验证了UTC-tree的有效性以及DPG查询的可扩展性。

关键词: 时序地理社交网络, 动态偏好, 查询索引, 组查询, 时间窗口