Analysis And Prediction Of Metropolitan Rail Passenger Volume Before Holidays

When people are enjoying the city’s fashionable and convenient in the process ofrapid urbanization, they also meet with traffic congestion, environmental pollutionand other urban disease.Mega-cities like Beijing, with the continuous development of economy, urbanpopulation further increased, and the vehicle population soared, traffic problems havebecome more serious, especially in the New Year’s Day, Tomb Sweeping Day, LaborDay, Dragon Boat Festival, Mid-Autumn Festival, National Day and other holidaysapproaching, People are suffering from more congested travel. The arrival of theholidays presents a challenge for a city transportation system, especially for publictransport. Therefore, it is necessary to analysis and forecast the public transportpassenger volume characteristics before holidays for helping to manage the traffic.Firstly, this paper determined the rail transport passenger volume as a researchsubject from the analysis of the evolution of public transport development in Beijing,and Meanwhile classifying legal holidays into "3" day holiday and "7" day holiday.Then, according to analyze the factors and characteristics of rail transportpassenger volume on holidays, methods that suitable for the paper’s problem werefunded. It used locally weighted regression, gray GM (1,1) model, multiple linearregression and neural network model to predict rail transport passenger volume beforethe holiday, and finally selected integrated use of locally weighted regression andmultiple linear regression model as the prediction through comparative analysis of themerits of the four methods.It can provide more scientific and rational basis for traffic operation before theholidays by predicting travel peak point and passenger volume of rail transport beforeholidays, as well as to lay the foundation for the development of an active andeffective transport organization scheme that protecting the city traffic runningsmoothly.