Visual Analysis Of Potential Patterns Of Spatiotemporal Trajectory Data

With the rapid development of wireless communication and mobile computing technologies and accuracy of global positioning and navigation systems,and the amount of spatiotemporal trajectory data is increasing,so the demand for technologies capable of analyzing the potential modes of trajectories are also increasing.Moreover,real-world spatiotemporal phenomena present order and hierarchical nature.Pattern mining of spatio-temporal trajectory data is to discover some valuable potential modes,such as frequent mode,accompanying mode,aggregation mode,and abnormal mode.Traditional trajectory pattern mining researches focus on how to find a pattern more accurately and quickly.However,same trajectory sequence contains several different pattern features,the existing method can not easily find the behavioral patterns of the moving object.The process of pattern evolution over time.In addition,the work of visual analysis and the space of exploration have certain limitations.Aiming at the phenomenon that trajectory sequences may exhibit different behavior patterns at different time granularities,this paper proposes a visual analysis method based on trajectory clustering for subject model mining,which is made up by combining trajectory data pattern mining and visualization technology.In particular,in this paper,a trajectory pattern analysis algorithm based on Partition-and-Group model is suggested,and proposed a pattern mining method based on dynamic time-bending distance after clustering.Furthermore,we designed a multi-time granularity trajectory pattern analysis view,and established a visual interaction analysis model of potential patterns.Users can change the time partition to check the change of the subject mode in the trajectory data with time.In addition,in order to demonstrate its function and practical value,our research uses the Stark experimental project animal mobile dataset to verify the method.The experimental results show that the proposed method can first provide a highly customizable visual interaction research platform for users to carry out exploratory analysis tasks.Secondly,it can flexibly mine and analyze the change of the subject mode in the trajectory data.