Non-parametric machine learning methods for evaluating the effects of traffic accident duration on freeways

Ying Lee, Chien-Hung Wei, Kai Chon Chao

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

Traffic accidents usually cause congestion and increase travel-times. The cost of extra travel time and fuel consumption due to congestion is huge. Traffic operators and drivers expect an accurately forecasted accident duration to reduce uncertainty and to enable the implementation of appropriate strategies. This study demonstrates two non-parametric machine learning methods, namely the k-nearest neighbour method and artificial neural network method, to construct accident duration prediction models. The factors influencing the occurrence of accidents are numerous and complex. To capture this phenomenon and improve the performance of accident duration prediction, the models incorporated various data including accident characteristics, traffic data, illumination, weather conditions, and road geometry characteristics. All raw data are collected from two public agencies and were integrated and cross-checked. Before model development, a correlation analysis was performed to reduce the scale of interrelated features or variables. Based on the performance comparison results, an artificial neural network model can provide good and reasonable prediction for accident duration with mean absolute percentage error values less than 30%, which are better than the prediction results of a k-nearest neighbour model. Based on comparison results for circumstances, the Model which incorporated significant variables and employed the ANN method can provide a more accurate prediction of accident duration when the circumstances involved the day time or drunk driving than those that involved night time and did not involve drunk driving. Empirical evaluation results reveal that significant variables possess a major influence on accident duration prediction.

Original languageEnglish
Pages (from-to)91-104
Number of pages14
JournalArchives of Transport
Volume43
Issue number3
DOIs
Publication statusPublished - 2017 Jan 1

Fingerprint

Highway accidents
Highway systems
traffic accident
learning method
Learning systems
Accidents
accident
Travel time
neural network
travel
traffic
Neural networks
performance comparison
development model
Fuel consumption
Lighting
driver
mathematics
uncertainty
road

All Science Journal Classification (ASJC) codes

  • Automotive Engineering
  • Transportation

Cite this

@article{c39b631a54a44415b232d1d8c8816ccd,
title = "Non-parametric machine learning methods for evaluating the effects of traffic accident duration on freeways",
abstract = "Traffic accidents usually cause congestion and increase travel-times. The cost of extra travel time and fuel consumption due to congestion is huge. Traffic operators and drivers expect an accurately forecasted accident duration to reduce uncertainty and to enable the implementation of appropriate strategies. This study demonstrates two non-parametric machine learning methods, namely the k-nearest neighbour method and artificial neural network method, to construct accident duration prediction models. The factors influencing the occurrence of accidents are numerous and complex. To capture this phenomenon and improve the performance of accident duration prediction, the models incorporated various data including accident characteristics, traffic data, illumination, weather conditions, and road geometry characteristics. All raw data are collected from two public agencies and were integrated and cross-checked. Before model development, a correlation analysis was performed to reduce the scale of interrelated features or variables. Based on the performance comparison results, an artificial neural network model can provide good and reasonable prediction for accident duration with mean absolute percentage error values less than 30{\%}, which are better than the prediction results of a k-nearest neighbour model. Based on comparison results for circumstances, the Model which incorporated significant variables and employed the ANN method can provide a more accurate prediction of accident duration when the circumstances involved the day time or drunk driving than those that involved night time and did not involve drunk driving. Empirical evaluation results reveal that significant variables possess a major influence on accident duration prediction.",
author = "Ying Lee and Chien-Hung Wei and Chao, {Kai Chon}",
year = "2017",
month = "1",
day = "1",
doi = "10.5604/01.3001.0010.4228",
language = "English",
volume = "43",
pages = "91--104",
journal = "Archives of Transport",
issn = "0866-9546",
publisher = "Versita",
number = "3",

}

Non-parametric machine learning methods for evaluating the effects of traffic accident duration on freeways. / Lee, Ying; Wei, Chien-Hung; Chao, Kai Chon.

In: Archives of Transport, Vol. 43, No. 3, 01.01.2017, p. 91-104.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Non-parametric machine learning methods for evaluating the effects of traffic accident duration on freeways

AU - Lee, Ying

AU - Wei, Chien-Hung

AU - Chao, Kai Chon

PY - 2017/1/1

Y1 - 2017/1/1

N2 - Traffic accidents usually cause congestion and increase travel-times. The cost of extra travel time and fuel consumption due to congestion is huge. Traffic operators and drivers expect an accurately forecasted accident duration to reduce uncertainty and to enable the implementation of appropriate strategies. This study demonstrates two non-parametric machine learning methods, namely the k-nearest neighbour method and artificial neural network method, to construct accident duration prediction models. The factors influencing the occurrence of accidents are numerous and complex. To capture this phenomenon and improve the performance of accident duration prediction, the models incorporated various data including accident characteristics, traffic data, illumination, weather conditions, and road geometry characteristics. All raw data are collected from two public agencies and were integrated and cross-checked. Before model development, a correlation analysis was performed to reduce the scale of interrelated features or variables. Based on the performance comparison results, an artificial neural network model can provide good and reasonable prediction for accident duration with mean absolute percentage error values less than 30%, which are better than the prediction results of a k-nearest neighbour model. Based on comparison results for circumstances, the Model which incorporated significant variables and employed the ANN method can provide a more accurate prediction of accident duration when the circumstances involved the day time or drunk driving than those that involved night time and did not involve drunk driving. Empirical evaluation results reveal that significant variables possess a major influence on accident duration prediction.

AB - Traffic accidents usually cause congestion and increase travel-times. The cost of extra travel time and fuel consumption due to congestion is huge. Traffic operators and drivers expect an accurately forecasted accident duration to reduce uncertainty and to enable the implementation of appropriate strategies. This study demonstrates two non-parametric machine learning methods, namely the k-nearest neighbour method and artificial neural network method, to construct accident duration prediction models. The factors influencing the occurrence of accidents are numerous and complex. To capture this phenomenon and improve the performance of accident duration prediction, the models incorporated various data including accident characteristics, traffic data, illumination, weather conditions, and road geometry characteristics. All raw data are collected from two public agencies and were integrated and cross-checked. Before model development, a correlation analysis was performed to reduce the scale of interrelated features or variables. Based on the performance comparison results, an artificial neural network model can provide good and reasonable prediction for accident duration with mean absolute percentage error values less than 30%, which are better than the prediction results of a k-nearest neighbour model. Based on comparison results for circumstances, the Model which incorporated significant variables and employed the ANN method can provide a more accurate prediction of accident duration when the circumstances involved the day time or drunk driving than those that involved night time and did not involve drunk driving. Empirical evaluation results reveal that significant variables possess a major influence on accident duration prediction.

UR - http://www.scopus.com/inward/record.url?scp=85030469948&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85030469948&partnerID=8YFLogxK

U2 - 10.5604/01.3001.0010.4228

DO - 10.5604/01.3001.0010.4228

M3 - Article

AN - SCOPUS:85030469948

VL - 43

SP - 91

EP - 104

JO - Archives of Transport

JF - Archives of Transport

SN - 0866-9546

IS - 3

ER -