Latent Class Analysis and Severity Model Evaluation for Vehicle-to-Vehicle Crashes in Urban Areas

Woowon Kim; Sungjun Lee; Juneyoung Park; Junhan Cho

doi:10.7470/jkst.2025.43.6.716

All Issue

2025 Vol.43, Issue 6 Preview Page Next Page

Article

Latent Class Analysis and Severity Model Evaluation for Vehicle-to-Vehicle Crashes in Urban Areas 도심부 차대차 사고 잠재계층분석 및 심각도 모형 평가

31 December 2025. pp. 716-731

PDF XML

Abstract

With the rapid growth of cities and increasing vehicle penetration, safety in urban areas has become an important social issue. In particular, 4-5 lane section shows a complex pattern due to various road components such as intersections, and tends to increase the severity of crashes due to high driving speeds. In addition, various factors such as increased traffic volume during peak hours and frequent lane changes due to staggered sections increase the risk of crashes. Considering these complex factors, this study aims to analyze the main factors influencing the severity of 4-5 lane urban road crashes. A total of 212 vehicle-to-vehicle crash cases were collected from dashcam footage on 4-5 lane urban dual carriageways across the country in 2021. Latent Class Analysis was conducted to capture latent heterogeneity in crash patterns. Additionally, to address data shortage, the dataset was augmented using the Random Over-Sampling Examples method. Subsequently, both the Binary Logit Model and the Random Parameter Binary Logit Model were estimated to analyze the factors influencing crash severity. The analysis identified two primary latent classes: Class 1 was dominated by rear-end crashes occurring on snowy or wet road surfaces, while Class 2 was characterized by side-impact crashes occurring under clear weather and dry surface conditions. Key determinants of crash severity included type of crash, road surface condition, and time period. Notably, both models indicated that rainy weather conditions were associated with elevated predicted crash severity, and wet and snowy road surfaces also showed significant associations with increased severity. Furthermore, the Random Parameter model outperformed the fixed-effect model across all performance metrics indicating that accounting for unobserved class yields more effective insights into crash severity. These findings are expected to contribute to the development of targeted safety strategies aimed at improving the safety of urban 4-5 lane roadways.

Keywords

crash severity

latent class analysis

random over-sampling examples

random parameter binary logit model

urban area

도시의 급속한 성장과 차량 보급의 증가로 인해 도심부 교통안전은 중요한 사회적 과제로 대두되고 있다. 특히 왕복 4~5차로 도로는 진출입부, 교차로 등 다양한 요소가 밀집되어 있으며, 높은 주행 속도와 빈번한 차로 변경으로 인해 사고 발생 시 심각도가 높아질 가능성이 크다. 본 연구는 도심부 왕복 4~5차로에서 발생한 차대차 사고를 대상으로 사고 심각도에 영향을 미치는 주요 요인을 식별하고 이질적 특성을 고려한 분석을 목적으로 한다. 이를 위해 2021년 전국 도심부 왕복 4~5차로 구간에서 발생한 교통사고 중 블랙박스 영상으로 수집된 212건의 사고 데이터를 분석에 활용하였다. 사고의 잠재적 이질성을 반영하기 위해 잠재계층분석을 수행하였으며, 데이터의 부족 문제를 완화하기 위해 Random Over-Sampling Examples 기법으로 데이터를 증강하였다. 이후 이항 로짓 모델과 랜덤 파라미터 이항 로짓 모델을 적용하여 사고 심각도에 미치는 영향 요인을 분석하였다. 분석 결과, 사고 데이터는 두 개의 주요 계층으로 분류되었으며, 계층 1은 적설 및 젖은 노면 상태에서 발생한 후미추돌 사고가 지배적인 특성을 보였고, 계층 2는 맑은 날씨와 건조한 노면에서 발생한 측면 추돌이 분포하는 특성을 나타냈다. 심각도에 영향을 미치는 주요 요인은 사고 유형, 노면 상태, 시간대 등으로 도출되었으며, 특히 두 모델 모두 비가 오는 기상 조건에서 높은 사고 심각도 추정값이 도출되었다. 또한 노면이 습윤 및 적설 상태 역시 심각도 증가와 유의한 관련성을 보였다. 랜덤 파라미터 모델이 여러 평가지표에서 우수한 성능을 보임에 따라, 개별 사고의 이질성을 반영한 방식이 사고 심각도 분석에 효과적임을 알 수 있었다. 이러한 결과를 통해 도심부 왕복 4~5차로의 안전성 향상을 기대할 수 있다.

키워드

사고 심각도

잠재계층분석

데이터 샘플링

랜덤 파라미터 이항 로짓 모델

도심부

References

Abdel-Aty M. (2003), Analysis of Driver Injury Severity Levels at Multiple Locations using Ordered Probit Models, Journal of Safety Research, 34(5), 597-603.

10.1016/j.jsr.2003.05.009

Chen J., Pu Z., Zheng N., Wen X., Ding H., Guo X. (2024), A Novel Generative Adversarial Network for Improving Crash Severity Modeling with Imbalanced Data, Transportation Research Part C: Emerging Technologies, 164, 104642.

10.1016/j.trc.2024.104642

Chen M., Zhou L., Choo S., Lee H. (2022), Analysis of Risk Factors Affecting Urban Truck Traffic Accident Severity in Korea, Sustainability, 14(5), 2901.

10.3390/su14052901

Cho J. H., Lee S. J., Park S. M., Park J. Y. (2022), Classification and Prediction of Highway Accident Characteristics Using Vehicle Black Box Data, The Journal of The Korea Institute of Intelligent Transport Systems, 21(6), 132-145.

10.12815/kits.2022.21.6.132

Depaire B., Wets G., Vanhoof K. (2008), Traffic Accident Segmentation by Means of Latent Class Clustering, Accident Analysis & Prevention, 40(4), 1257-1266.

10.1016/j.aap.2008.01.007

Fountas G., Sarwar M. T., Anastasopoulos P. C., Blatt A., Majka K. (2018), Analysis of Stationary and Dynamic Factors Affecting Highway Accident Occurrence: A Dynamic Correlated Grouped Random Parameters Binary Logit Approach, Accident Analysis & Prevention, 113, 330-340.

10.1016/j.aap.2017.05.018

Jo Y., Jung A. R., Park H. J., Park J. H., Oh C. (2022), Prioritizing Driving Safety Indicators using Real-World C-ITS Data to Identify Hazardous Freeway Sections, J. Korean Soc. Transp., 40(6), Korean Society of Transportation, 863-878.

10.7470/jkst.2022.40.6.863

Kim H. S., Park N. R., Park J. Y., Wang L. (2023), Accident Analysis for Port Terminals using Heterogeneous XAI-Informed Machine Learning Approach, In 2023 IEEE International Conference on Knowledge Graph (ICKG), 227-234, IEEE.

10.1109/ICKG59574.2023.00033

Kim J. K., Al-Bdairi N. S. S., Hernandez S. (2024), Lane Number and Its Impact on Commercial Motor Vehicle Crash Safety: An Econometric Perspective, International Journal of Transportation Science and Technology, 16, 119-133.

10.1016/j.ijtst.2024.01.002

Kim S. G. (2021), Analysis of Exposure Variables for Determining Traffic Accident Risk Indicators and Equal International Comparison, J. Korean Soc. Transp., 39(6), Korean Society of Transportation, 780-792.

10.7470/jkst.2021.39.6.780

Ko C. W., Kim H. M., Jeong Y. S., Kim J. H. (2020), A Study on Injury Severity Prediction for Car-to-Car Traffic Accidents, The Journal of The Korea Institute of Intelligent Transport Systems, 19(4), 13-29.

10.12815/kits.2020.19.4.13

Ko J. E., Jang J. Y., Oh C. (2021), Assessing the Safety Benefits of In-Vehicle Warning Information by Vehicle Interaction Analysis in C-ITS Environments, J. Korean Soc. Transp., 39(1), Korean Society of Transportation, 1-13.

10.7470/jkst.2021.39.1.001

Korea Transportation Safety Authority (2017), Development planning of integrated evaluation technology for vehicle-to-vehicle collision safety based on real accidents (실사고 기반 차대차 충돌안전성 통합 평가기술 개발 기획).

Lee S.H. , Park J.Y. (2024), A Study on Autonomous Vehicle Crash Hierarchy Analysis and Severity Model Based on Bayesian Probabilistic Inference, J. Korean Soc. Transp., 42(1), Korean Society of Transportation, 77-93.

10.7470/jkst.2024.42.1.077

Li Z., Wu Q., Ci Y., Chen C., Chen X., Zhang G. (2019), Using Latent Class Analysis and Mixed Logit Model to Explore Risk Factors on Driver Injury Severity in Single-Vehicle Crashes, Accident Analysis and Prevention, 129, 230-240.

10.1016/j.aap.2019.04.001

Man C. K. (2022), Developing Transferable Real-Time Crash Prediction Models for Highly Imbalanced Data (Doctoral dissertation, Loughborough University).

Menardi G., Torelli N. (2014), Training and Assessing Classification Rules with Imbalanced Data, Data mining and knowledge discovery, 28(1), 92-122.

10.1007/s10618-012-0295-5

Park S. J., Kho S. Y., Park H. C. (2019), The Effects of Road Geometry on the Injury Severity of Expressway Traffic Accident depending on Weather Conditions, The Journal of The Korea Institute of Intelligent Transport Systems, 18(2), 12-28.

10.12815/kits.2019.18.2.12

Police Science Institute (2013), A study on the implementation plan of next-generation ITS (C-ITS) technology for traffic accident reduction (교통사고 감소를 위한 차세대 ITS(C-ITS) 기술 도입방안 연구).

Redelmeier D. A., Raza S. (2017), Life-Threatening Motor Vehicle Crashes in Bright Sunlight, Medicine, 96(1), e5710.

10.1097/MD.000000000000571028072708PMC5228668

Seiniger P., Bartels O., Pastor C., Wisch M. (2013), An Open Simulation Approach to Identify Chances and Limitations for Vulnerable Road User (VRU) Active Safety, Traffic injury prevention, 14(sup1), S2-S12.

10.1080/15389588.2013.797574

Tefft B. C. (2016), Motor Vehicle Crashes, Injuries, and Deaths in Relation to Weather Conditions, United States, 2010-2014.

Won M. S., Lee G. R., Gang G. U. (2009), A Study on the Application of Accident Severity Prediction Model, J. Korean Soc. Transp., 27(4), Korean Society of Transportation, 167-173.

Wu Q., Chen F., Zhang G., Liu X. C., Wang H. Bogus S. M.(2014), Mixed Logit Model-Based Driver Injury Severity Investigations in Single-and Multi-Vehicle Crashes on Rural Two-Lane Highways, Accident Analysis & Prevention, 72, 105-115.

10.1016/j.aap.2014.06.014

Zarei M., Hellinga B. (2021), Crash Data Augmentation using Conditional Generative Adversarial Networks (CGAN) for Improving Safety Performance Functions, arXiv preprint arXiv:2112.12263.

Information

Publisher :Korean Society of Transportation
Publisher(Ko) :대한교통학회
Journal Title :Journal of Korean Society of Transportation
Journal Title(Ko) :대한교통학회지
Volume : 43
No :6
Pages :716-731
Received Date : 2025-06-20
Revised Date : 2025-07-18
Accepted Date : 2025-09-23
DOI :https://doi.org/10.7470/jkst.2025.43.6.716

Journal of Korean Society of Transportation ISSN:1229-1366(Print) 2234-4217(Online) 대한교통학회지

All Issue