Modeling and Sensitivity Analysis for Railway Passengers’ Public Transit Mode Choice

Abstract

Comprehensive terminal integrates different modes of transport and sharing rates of different traffic modes influence its planning greatly. To determine the sharing rates, it’s necessary to study mode choice probability of passengers. Based on the theory of disaggregate choice, with the survey of mode choices of passengers arriving by rail in Nanjing south railway station, this paper establishes multinomial logit (MNL) model and nest logit (NL) model, considering three types of modes of transport including the subway, regular bus, and taxi. Models are calibrated with SPSS and validated through statistical indicators. Therefore, important factors influencing choice are obtained. The accuracy test results of two kinds of models show that NL model has higher prediction accuracy, and can overcome the shortcomings of traditional logit model’s independent and irrelevant alternatives (IIA). Finally, this paper develops the sensitivity analysis using the NL upper model as an example, and gives the sensitivity physical meaning, namely the development level of taxi.