Comparison of ANN and Multinomial Logit for Prediction of Mode Choice between Online Transportations and Private Vehicles
Puguh Budi Prakoso, Yuslena Sari, Iphan Fitrian Radam, Najmi Fahrina
Faculty of Engineering, Universitas Lambung Mangkurat, Indonesia
National Center for Sustainable Transportation Technology, Indonesia
e-mail: puguh.prakoso[at]ulm.ac.id, ifradam[at]ulm.ac.id, najmi.fahrina[at]gmail.com
Abstract
The presence of online transportation modes is a manifestation of the rapid development of information technology, especially smartphone and apps technology. Because of the critical issues of safety, service quality assurance and effectiveness of load factor to its capacity, online transportations in Indonesia cannot be classified and regulated as public transportation. The main function of online transportation is actually as a short distance feeder to public transportation. Unfortunately, because the public transportation is not adequate, so the online transportation functions slowly become informal public transportation mode. This in the future can cause problems in a city, especially when the increasing number of online transportations will rise the vehicle traffic together with private vehicles. To understand and prevent this problem becoming bigger in the near future, the mode choice between private vehicles and online transportations requires to be studied first. The purpose of this research is to obtain a reliable model of modal choice between private vehicles and online transportation by comparing the multinomial logit (MNL) probability method with the Artificial Neural Network (ANN), to study the accuracy of the performance of the two models for the mode selection between private vehicles and online transportation, and to evaluate the socio-economic factors underlying the mode choice between private vehicles and online transportation in the city areas of Banjarmasin and Banjarbaru.
Keywords: mode choice, multinomial logit probability method, ANN, online transportation
Topic: EV Socio-Economical Impact