Research on Passenger Flow Prediction of Beijing Subway Based on Spatiotemporal Correlation Analysis

Urban rail transit has significant advantages such as large traffic volume, fast speed and high comprehensive benefits. It has become the most important component of urban traffic construction management and urban traffic congestion solution. Based on the actual passenger flow data of the Beijing subway Line 2 within one day, this paper analyzes the unbalanced distribution of passenger flow in space and time by using the method and theory of short-term traffic flow prediction. The gray correlation between the overall passenger flow of Line 2 and the passenger flow data of each site in one day was compared, and four stations with the highest gray correlation were selected for subsequent passenger flow forecasting. Then, based on the passenger flow of Line 2, four stations and corresponding total passenger flow data are taken as training samples, and the time is from 4:00 to 18:00. Taking the data of the four stations with the highest correlation as input, the short-term prediction of the passenger flow from 18:00 to 19:00 is used as the output.