A novel micro individual behavior characterization method based on trajectory big data

The spatiotemporal trajectory of population is a key factor affecting energy consumption behaviors and levels. Mining micro individual behavior patterns based on trajectory data and finely depicting them is a reliable means of accurately and comprehensively characterizing the flexible energy use potential, spatiotemporal distribution, and response characteristics of the population. Aiming at the problem of insufficient representation of individual behavior by existing methods, a trajectory data hierarchical processing flow and a micro individual behavior representation method based on physical information constraints driven by trajectory big data are proposed. Firstly, due to the presence of a large amount of noise in trajectory data that reduces computational accuracy and redundant data that reduces computational efficiency, a hierarchical processing method for trajectory data is proposed to improve the quality of trajectory data; Secondly, build an OPTICS clustering model based on physical information constraints such as speed to achieve accurate identification of core stationary areas in both moving and stationary states. Finally, a case study was conducted on the Geolife Trajectory 1.3 dataset, and the results showed that the proposed method can achieve trajectory data cleaning. The average error of the processed trajectory data was 16.86 meters, and the trajectory data volume was compressed to 30% of the original data volume. At the same time, the proposed method accurately characterizes the multi temporal and spatial scale activity behavior of microscopic individuals: there are 9 core stationary regions, with the stationary state accounting for 42.6% of all trajectory data and the motion state accounting for 57.4%. This work provides a solid foundation for the analysis and identification of flexible energy consumption patterns.