Impact of risk factors on work zone crashes using logistic models and Random Forest

Thumbnail Image

Files

2104.06561.pdf (605.33 KB)

Links to Files

https://ieeexplore.ieee.org/abstract/document/9564405

Permanent Link

http://hdl.handle.net/11603/23039
 https://doi.org/10.1109/ITSC48978.2021.9564405

Collections

UMBC Data Science
UMBC Faculty Collection

Author/Creator

Author/Creator ORCID

https://orcid.org/0000-0002-6835-8338

Date

2021-04-14

Type of Work

Conference papers and proceedings
preprints
Text

Department

Program

Citation of Original Publication

H. I. Ashqar, Q. H. Q. Shaheen, S. A. Ashur and H. A. Rakha, "Impact of risk factors on work zone crashes using logistic models and Random Forest," 2021 IEEE International Intelligent Transportation Systems Conference (ITSC), 2021, pp. 1815-1820, doi: 10.1109/ITSC48978.2021.9564405.

Rights

© 2021 IEEE.  Personal use of this material is permitted.  Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.

Subjects

Abstract

Work zone safety is influenced by many risk factors. Consequently, a comprehensive knowledge of the risk factors identified from crash data analysis becomes critical in reducing risk levels and preventing severe crashes in work zones. This study focuses on the 2016 severe crashes that occurred in the State of Michigan (USA) in work zones along highway I-94. The study identified the risk factors from a wide range of crash variables characterizing environmental, driver, crash and road-related variables. The impact of these risk factors on crash severity was investigated using frequency analyses, logistic regression statistics, and a machine learning Random Forest (RF) algorithm. It is anticipated that the findings of this study will help traffic engineers and departments of transportation in developing work zone countermeasures to improve safety and reduce the crash risk. It was found that some of these factors could be overlooked when designing and devising work zone traffic control plans. Results indicate, for example, the need for appropriate traffic control mechanisms such as harmonizing the speed of vehicles before approaching work zones, the need to provide illumination at specific locations of the work zone, and the need to establish frequent public education programs, flyers, and ads targeting high-risk driver groups. Moreover, the Random Forest algorithm was found to be efficient, promising, and recommended in crash data analysis, specifically, when the data sample size is small.