Prediction of Road Network Speed around Sports Events Based on Random Forest Model

With national economic improvements, more large-scale activities are held in major cities, causing pressure on surrounding road operation. Therefore, this paper studies the average speed of surrounding roads during short-term large-scale activities and analyzes temporal and spatial characteristics of their impact on surrounding roads using floating car and coil detector data. Analysis shows that average speed of the affected road around the venue declines and traffic volume of the section increases significantly two hours before the activity, and average speed of the road decreases and traffic volume increases within 1 hour after the activity. The prediction results show that the average 15-minute speed prediction accuracy of East Second Ring Road, New East Road, North Workers Stadium Road, and East Workers Stadium Road are 94.09%, 88.90%, 78.11% and 84.64% respectively. The prediction results can provide corresponding support for the study and judgment of surrounding road network operation situation during large-scale activities.