云计算环境下顾及用户关系的手机用户时空轨迹模式挖掘方法研究

项目名称： 云计算环境下顾及用户关系的手机用户时空轨迹模式挖掘方法研究

项目编号： No.41471371

项目类型： 面上项目

立项/批准年度： 2015

项目学科： 天文学、地球科学

项目作者： 吉根林

作者单位： 南京师范大学

项目金额： 90万元

中文摘要： 随着GPS定位和无线通信技术的发展，人们不仅可以记录手机的通话信息，而且可以定位手机所在位置，采样用户时空轨迹数据。从手机用户的时空轨迹数据和通话记录中挖掘用户轨迹模式与用户关系，对于用户行为模式分析、社交网络分析、热点地区发现、给用户提供位置服务与个性化推荐等具有重要的理论意义和应用价值。由于手机用户时空轨迹和通话记录是大数据，因此本课题以Map/Reduce为并行计算模型，研究手机用户时空轨迹模式挖掘方法，包括用户频繁轨迹模式、异常轨迹模式、聚集模式（热点区域）和伴随模式；综合利用时空轨迹数据和通话记录建立用户亲近度模型，发现用户亲近关系和团伙关系，创新地提出多种手机用户时空轨迹模式和用户关系挖掘算法，以及相应并行算法，将所提挖掘算法在云环境下加以实现，并利用真实数据源进行实验与应用，分析算法性能。在云计算环境下同时挖掘手机用户时空轨迹模式与用户关系，尚未见文献报道。

中文关键词： 时空轨迹挖掘；用户关系挖掘；轨迹模式挖掘；时空数据挖掘；大数据

英文摘要： With the development of GPS positioning and wireless communication technology, people can not only record the mobile phone call information, but also can locate mobile phone location, sampling the user spatio-temporal trajectory. It has important theoretical significance and application value to mining user trajectory patterns and user relationships from spatio-temporal trajectory and call information for analysis of user behavior patterns and social network, finding hot area, providing position services and personalized recommendation. As spatio-temporal trajectory and call information of the mobile phone user are big data, so in this project, based on parallel computing model Map/Reduce, mining methods of spatio-temporal trajectory patterns of mobile phone user are researched. The spatio-temporal trajectory patterns include frequent trajectory pattern, abnormal trajectory pattern, gathering pattern and adjoint pattern. User close degree modle is created and users close relationship and group are discovered from spatio-temporal trajectory and call records. We will creatively put forward a variety of mining algorithms for spatio-temporal trajectory patterns and user relationship, and the corresponding parallel algorithms. The proposed algorithms will be implemented in a cloud environment. We will carry on experiments, algorithm performance analysis and applications using real data sets. Mining spatio-temporal trajectory patterns and user relations of mobile phone users under cloud computing environment has not been reported.

英文关键词： Spatio-temporal trajectory mining;User relationships mining;Trajectory Patterns mining;Spatio-temporal data mining;Big data