Short-Term Traffic Speed Prediction via Machine Learning

Authors

Luis Romo, The University of Texas Rio Grande Valley
Jingru Zhang
Kevin Eastin, The University of Texas Rio Grande Valley
Chao Xue

Document Type

Conference Proceeding

Publication Date

11-21-2020

Abstract

Short-term traffic speed prediction is a key component of Intelligent Transportation Systems (ITS), which has an impact on travelers’ routing decisions and behaviors to the traffic congestion. In the past years, traffic speed prediction has been studied a lot and different machine learning methods are employed, including deep learning approaches, which recently attracts much attention from both academic and industry fields. In this work, we investigate three different machine learning methods for predicting the short-term traffic speed, i.e., Convolutional Neural Network, Long Short-term Memory Neural Network and Extreme Gradient Boost. The training and testing data are collected by ourselves from the California Department of Transportation. Through comparisons with the baseline average method, it is obvious that machine learning approaches can achieve more accurate and stable prediction performance.

Comments

© 2020 Springer Nature Singapore Pte Ltd.

https://rdcu.be/eBgCt

Recommended Citation

Romo, Luis, Jingru Zhang, Kevin Eastin, and Chao Xue. "Short-term traffic speed prediction via machine learning." In International Conference on Green, Pervasive, and Cloud Computing, pp. 31-42. Singapore: Springer Singapore, 2020. https://doi.org/10.1007/978-981-33-4532-4_3

Publication Title

International Conference on Green, Pervasive, and Cloud Computing

DOI

10.1007/978-981-33-4532-4_3