Prediction System of Bus Arrival Time Based on Historical Data Using Regression Models

Authors

Kássio Ribeiro Coquita, Universidade Federal de Pernambuco (UFPE)Follow
Arley Ramalho Rodrigues Ristar, Universidade Federal de Pernambuco (UFPE)Follow
Adriano Lorena Inácio de Oliveira, Universidade Federal de Pernambuco (UFPE)Follow
Patricia Cabral de Azevedo Restelli Tedesco, Universidade Federal de Pernambuco (UFPE)Follow

Document Type

Article

Publication Date

5-2015

Keywords

Intelligent Transportation Systems, Prediction, Regression, SVR, ELM

Abstract

Intelligent Transportation Systems are applications of information and communication technologies aimed at improving the transportation area. Providing information about bus arrival time on the bus stop is very important and useful to passengers and transit managers. This paper presents a proposed bus arrival time prediction system at the bus stop where user is located. For this, an experimental study on bus route named Campina do Barreto No. 722 in Recife-PE, was performed by comparing the regression model for Support Vector Machine (SVR) and the neural network Extreme Learning Machine (ELM) to estimate the time it takes to go through adjacent bus stops. The experiments were performed using the bus GPS log data in the metropolitan region of Recife, and the results showed that for this application the SVR significantly outperforms ELM.

Comments

This paper is in Portuguese (Sistema de Previsão do Tempo de Chegada dos Ônibus Baseado em Dados Históricos Utilizando Modelos de Regressão)

Recommended Citation

Coquita, Kássio Ribeiro; Ristar, Arley Ramalho Rodrigues; de Oliveira, Adriano Lorena Inácio; and Tedesco, Patricia Cabral de Azevedo Restelli, "Prediction System of Bus Arrival Time Based on Historical Data Using Regression Models" (2015). Proceedings of the XI Brazilian Symposium on Information Systems (SBSI 2015). 84.
https://aisel.aisnet.org/sbis2015/84