Parameter Calibration of Traffic Flow Speed-Density Model Based on K-means Clustering Algorithm and Least Square Method

This paper originates from the objective to verify the effectiveness of the K-means clustering algorithm in dealing with large sample data deviation and better describes the characteristics of highway traffic flow. The data was downloaded from the open source data of traffic flow detector of Whitemud Drive Highway in Canada. After data cleaning, K-means Clustering Algorithm was used to cluster large sample data to solve the problem of data deviation. Then the parameter calibration of Greenshields and Underwood traffic flow speed-density models according to the training set after clustering was conducted based on Least Square Method. Results of the two models are evaluated by means of statistics and probability theory. Finally, evaluation results show that the fitting effect of the Greenshields model is better than the Underwood model, and it is more suitable for decribing the traffic flow state of the highway in the study section.