Citation

Li Y, Yamamoto T, Zhang G. J. Saf. Res. 2018; 64: 155-162.

Affiliation

Institute of Guangdong, Hong Kong and Macao Development Studies, Sun Yat-Sen University, Xingang Xi Road, Guangzhou, China; Center for Studies of Hong Kong, Macao and Pearl River Delta, Sun Yat-Sen University, Xingang Xi Road, Guangzhou, China. Electronic address: sysuzgn@gmail.com.

Copyright

(Copyright © 2018, U.S. National Safety Council, Publisher Elsevier Publishing)

DOI

10.1016/j.jsr.2017.12.002

PMID

29636164

Abstract

INTRODUCTION: Fatigue is one of the riskiest causes of traffic accidents threatening road safety. Due to lack of proper criteria, the identification of fatigue-related accidents by police officers largely depends on inferential evidence and their own experience. As a result, many fatigue-related accidents are misclassified and the harmfulness of fatigue on road safety is misestimated.

METHOD: In this paper, a joint model framework is introduced to analyze factors contributing to misclassification of a fatigue-related accident in police reports. Association rule data mining technique is employed to identify the potential interactions of factors, and logistic regression models are applied to analyze factors that hinder police officers' identification of fatigue-related accidents. Using the fatigue-related crash records from Guangdong Province during 2005-2014, factors contributing to the false positive and false negative detection of the fatigue-related accident have been identified and compared.

RESULTS: Some variables and interactions were identified to have significant impacts on fatigue-related accident detection.

CONCLUSIONS: Based on the results, it can be inferred that the stereotype of certain groups of drivers, crash types, and roadway conditions affects police officers' judgment on fatigue-related accidents. PRACTICAL APPLICATIONS: This finding can provide useful information for training police officers and build better criteria for fatigue identification.

Copyright © 2017 National Safety Council and Elsevier Ltd. All rights reserved.

Language: en

Keywords

Association rule; Fatigue; Logistic regression model; Misclassification; Police report