Linear Regression Model of Railway Passenger Traffic Volume and Its Influencing Factors

Abstract

The railway is one of the most crucial transportation modes in China and plenty of people choose to travel by railway. Finding the influencing factors of passenger traffic volume of the railway can help us to develop the railway construction. Substantial research has probed factors influencing the railway passenger volume and has identified several influencing factors. In this study, we try to find a long-term relationship between passenger traffic volume of the railway and six influencing factors, which are total population, employed persons, number of national railway passenger coaches owned, GDP, and household consumption expenditure. In order to find long term relationship models, we collected the data of r passenger traffic volume of the railway and other six influencing factors from 1978 to 2019. By developing several multiple linear regression models, we can see that all of these six factors can impact the passenger traffic volume of the railway and the model of six factors can describe the relationship well. After simplifying the model with six factors, we find two more models that can describe the relationship with less influencing factors, which are y = 1.059 × 105 - 1.773 × 10-2 x 1 + 2.729 × 10-1 x 5 + ε and log(y) = 11.5-3.309 × 10-8x1+1.435 × 10-6x5+ε. Using the diagnostic plots to analysis these two models, we find that log(y) = 11.5-3.309 × 10-8x1+1.435×10-6x5+ε is better. According to the model, GDP and traffic volume are two main influencing factors on the railway passenger volume. We use this model to predict the railway passengers in 2019 and compare the predict value with the data. The error rate i 5.4688% and we believe that this log(y) model can also be used to predict the railway passenger volume.