A method of trajectory stopping point extraction based on Gaussian model

Abstract

The invention discloses a trajectory stop point extraction method based on a Gaussian model. Given an initial radius and a density, the number of neighbor points within the radius range of each data point is calculated and recorded as the density of the point, and all trajectories whose density is less than MD are marked. Point, find the mean and variance of the number of neighbors of all marked trajectory points, establish the density Gaussian model of the moving point, adjust the density Gaussian model of the moving point correspondingly by adjusting the radius and density, until the structure of the remaining stopping points no longer changes, Obtain the final moving point density Gaussian model. The density Gaussian model is used to fit each trajectory data, and it is divided into different time periods to obtain the final aggregation pattern of stopping points. The method starts from the analysis of the aggregation characteristics of moving points, establishes a Gaussian model based on the density characteristics of moving points, and uses the Gaussian model to exclude moving points in the trajectory, thereby improving the extraction accuracy of stopping points.

Description

technical field [0001] The invention relates to a method for extracting a trajectory stopping point based on a Gaussian model, belonging to the technical field of trajectory data mining. Background technique [0002] Trajectory data is a data structure that describes the change of the moving position of a moving object over time, which contains a wealth of valuable information or knowledge. Usually, the trajectory data contains two types of data points: moving points and stopping points. The stopping points in the trajectory correspond to specific geographic locations or places, or locations where some important events occur, such as: supermarkets, shopping malls, schools, office buildings, gatherings and parades, traffic accidents, etc. Discovering these particular geographic locations can be used to analyze behavioral models of moving objects and predict when the next clustering pattern of that event will occur. Therefore, the stopping points in the trajectory are more i...

AI Technical Summary

Problems solved by technology

However, in the actual data distribution, the stop point reflects the complex behavior pattern of the moving object, and there are large differences between the speed, density, movement characteristics, and direction angle of each moving point, so that the extraction accuracy of the stop point is different on different trajectory data. a certain degree of volatility

In addition, in the same or different trajectories, the scales of the aggregation positions of stop points are different. For these stop points with different aggregation sizes, only using a uniquely determined neighborhood radius threshold greatly affects the accuracy of trajectory clustering.

Images