Quantitative Study on How Land Use Mix Impact Urban Rail Transit at Stationlevel

Abstract:

This essay focuses on how land use mix quantitatively impacts urban rail transit ridership at stationlevel by nonlinear regression model. Distancedecay weight and weight of population in mutual service area assigned to each station are used to weigh population within service area of station. Then ridership divided by weighted population is taken as dependent variable to analyze what is the relationship between land use mix and ridership at stationlevel. Least square support vector machine is the ideal model to do the above thing. Finally, data of 109 stations in Tokyo, Japan are taken as case study, result of which shows land use mix has a little influence on ridership at stationlevel and meanwhile, employment/inhabitants within service area of station has a significant influence on ridership at stationlevel. So, employment/inhabitants should substitute land use mix and be taken as key predictor for ridership at stationlevel.

Key words:traffic engineering; ridership forecasting at stationlevel; least squares support vector machine (LSSVM); land use mix; ratio of employment to inhabitant