Trajectory preprocessing method for taxi data set

Abstract

The invention discloses a trajectory preprocessing method for a taxi data set. The method comprises the steps of firstly obtaining trajectory data, wherein each sampling trajectory point comprises longitude, latitude and timestamp information; secondly analyzing the trajectory data, performing abstract storage in entity objects, numbering trajectories, and adding trajectory point IDs; thirdly searching for missing values of the trajectories, and supplementing the missing values by utilizing a linear interpolation method or a mean value method; fourthly clustering the trajectory points, detecting abnormal points, and accurately analyzing and processing the abnormal points; fifthly detecting a data redundancy region, extracting redundant data and performing trajectory compression; sixthly searching for the trajectory points at the corners in the trajectories, generating a corner point set, combining and adjusting the corner point set, and performing trajectory cutting according to the corner point set; and finally updating and outputting the trajectory information. According to the method, the missing values can be processed; the abnormal points can be detected and processed; and complex and overlapped trajectories can be cut.

Description

technical field [0001] The invention belongs to the technical field of data preprocessing, in particular to a method for trajectory preprocessing based on a taxi data set in a UCI machine learning library. Background technique [0002] With the rapid development of location acquisition technologies such as sensor networks, satellites, and wireless communications, various moving objects generate large-scale trajectory data. Trajectory data usually includes trajectory sequence and trajectory points, among which: trajectory point is the atomic data of the recorded trajectory, which is composed of longitude, latitude and time stamp; trajectory sequence is composed of several containing trajectory points. The data in real life is complicated and messy, and there are often missing and entry errors in the collected data. It is also a normal phenomenon to have missing values ​​in the trajectory data. If it is not processed, it will greatly interfere with the calculation results. Th...

AI Technical Summary

Problems solved by technology

In some areas of the trajectory data, the distribution of points is very dense and the characteristics are similar. If it is not processed, it will first lead to a large amount of data and a decrease in calculation speed. Secondly, it will also affect the accuracy of calculation.

Images