Abstract:

The acquisition of spatiotemporal trajectory data is becoming easier. Trajectory data depict the behavior patterns and activity rules of mobile objects. It is a true portrayal of the movement patterns and behavior characteristics of mobile objects in spatiotemporal environment. It has important application value in urban planning, traffic management, service recommendation, location prediction and other fields. These processes usually need to be realized through pattern mining of spatiotemporal trajectory data. Firstly, the preprocessing and basic steps of trajectory data mining are briefly described, and the classification of abnormal trajectory detection methods is summarized. Then, four kinds of trajectory pattern mining in recent years are analyzed and summarized. Finally, trajectory data mining is summarized and analyzed from the perspective of management decision, which is expected to provide necessary literature and theoretical basis for pattern mining and management decision-making of trajectory data.

Key words: trajectory data, trajectory preprocessing, spatiotemporal pattern mining, management decision

摘要：

时空轨迹数据的获取变得越来越容易，轨迹数据刻画了移动对象的行为模式与活动规律，是对移动对象在时空环境下的移动模式和行为特征的真实写照，在城市规划、交通管理、服务推荐、位置预测等领域具有重要的应用价值。这些过程通常需要通过对时空轨迹数据进行模式挖掘才能得以实现。简述了轨迹数据挖掘的预处理和基本步骤，归纳了异常轨迹检测方法的分类，分析、总结了近年来基于轨迹数据的四种模式挖掘，从管理决策角度对轨迹数据挖掘进行相关综述和分析，有望为轨迹数据的模式挖掘与管理决策提供必要的文献资料和理论基础。

关键词: 轨迹数据, 轨迹预处理, 时空模式挖掘, 管理决策