Trajectory Stops Identification Algorithm Based On Correlation Coefficient

With the rapid popularization of the modern mobile intelligent terminal, the improvement of the precision of track data and the rapid development of wireless communication system, It provides convenience for the acquisition and accurate analysis of the moving track. In real life, a large number of trajectory data is generated every day, due to the irregularity of the moving track and the low density of knowledge distribution, the difficulty in extracting and analyzing information is improving, and it brings more challenges to the trajectory research and puts forward higher requirements on the performance of related algorithms. In the aspect of trajectory analysis, how to extract the stay points in the trajectory accurately plays an important role in the field of trajectory identification, location prediction, and service recommendation.The paper, on the basis of analyzing the characteristics of real-time mobile trajectory and related research, according to irregularity of activities when people staying at stay points, combined with relationship between the degree of linear correlation of trajectory coordinates and the correlation coefficient, presents a algorithm of extracting trajectory stay points based on correlation coefficient. At the stay point, the track density is larger, and the track direction changes more frequently, due to that the correlation coefficient of the trace point coordinates can reflect the degree of the change of the trajectory direction and the density of the trajectory points at a certain extent, so it has important significance to discover and analyze the relevant key points before seeking stay points. The algorithm, considering the change of trajectory direction, the density and the residence time of the stay point as basic starting point, first filters the trajectory twice through correlation coefficient of trajectory point coordinate, generates key point sequence, locates potential stay points, then makes a comprehensive judgment on the potential stay points according to dwell time and density of key points related area, finally identifies the stay points in the trajectory. The method takes the key points as the core when identifying the stay points, and reduces the search scope and enhances the reliability of the recognition results. In order to improve the accuracy of the recognition region, adjacent points of the key points are also identified when judging and extracting the stay points. Finally, the feasibility and effectiveness of the algorithm are verified by experimental analysis.