详情

原文传递 Prediction of Traffic Incident Duration Using Clustering-Based Ensemble Learning Method

题名：	Prediction of Traffic Incident Duration Using Clustering-Based Ensemble Learning Method
正文语种：	eng
作者：	Hui Zhao;Willy Gunardi;Yang Liu;Christabel Kiew;Teck-Hou Teng;Xiao Bo Yang
作者单位：	Dept. of Industrial Systems Engineering and Management National Univ. of Singapore Block El #08-20 1 Engineering Dr. 2 Singapore 117576;Corporate Strategy and Development Cargill Inc. 138 Market St. #17-01 Capitagreen Singapore 048946;Dept. of Civil and Environmental Engineering and Dept. of Industrial Systems Engineering and Management National Univ.of Singapore Block E1A #06-09 1 Engineering Dr. 2 Singapore 117576;Dept. of Civil and Environmental Engineering National Univ. of Singapore Block E1 #08-20 1 Engineering Dr. 2 Singapore 117576;Data Analytics-Strategic Technology Centre Group Technology Office Singapore Technologies Engineering iHQ Pte. Ltd. 600 West Camp Rd. #01-01 Singapore 797654;Urban Solutions Singapore Technologies Engineering 100 Jurong East St. 21 Singapore 609602
关键词：	Traffic accidents; Incident duration prediction; Clustering model; Ensemble learning; Random forest (RF); Neural network
摘要：	Traffic incidents are a primary cause of traffic delays, which can cause severe economic losses. Effective traffic incident management requires integrating intelligent traffic systems, information dissemination, and the accurate prediction of incident duration. This study develops a clustering-based machine learning model to predict the incident duration. Unlike similar studies that train separate machine learning models for a fixed number of clusters, this study proposes an ensemble learning method based on multiple clustered individual models that can provide good and diverse prediction performance. The K-means clustering method is used in this study as a bootstrapping technique in the ensemble learning approach, with the individual models based on the artificial neural network model and random forest regression model. The models are tested using the incident data from Singapore, and the results show that the ensemble model outperforms both the traditional model with fixed clusters and the classical model without clustering. Additionally, this study attempted to determine the significance of different variables on traffic incident durations using the random forest feature importance function. The prediction of incident duration and the analysis of influence factors can contribute to several aspects of traffic management, such as improving traffic dissemination to mitigate traffic congestion caused by incidents.
出版年：	2022
期刊名称：	Journal of Transportation Engineering
卷：	148
期：	7
页码：	04022044.1-04022044.8