Journal of Transportation Research

Journal of Transportation Research

Data-Driven Modeling of Factors Affecting Speed Variance on Rural Roads Using a Machine Learning Approach (Case Study: Hamadan Province)

Document Type : Original Article

Authors
Ali Tavakoli Kashani 1
Mohammad Najafpour Derav 2
Amirreza Eskandari 3
Majid Najafi 3
1 Associate Professor, School of Civil Engineering, Iran University of Science and Technology, Tehran, Iran and Road Safety Research Center, Iran University of Science and Technology, Tehran, Iran.
2 Ph.D., Stud., School of Civil Engineering, Iran University of Science and Technology, Tehran, Iran and Road Safety Research Center, Iran University of Science and Technology, Tehran, Iran.
3 M.Sc., Student, School of Civil Engineering, Iran University of Science and Technology, Tehran, Iran.
10.22034/tri.2025.556481.3397
Abstract
Speed variance is widely acknowledged as a critical indicator of traffic flow stability and a major determinant of road safety, as excessive fluctuations in vehicle speed significantly elevate the likelihood and severity of crashes. This study aims to identify and model the key factors influencing speed variance on rural highways in Iran using the Random Forest algorithm—a robust machine learning method capable of capturing complex nonlinear relationships and reducing overfitting risk.

A total of 200 observation points were collected from various segments of the rural road network through detailed field surveys. The dataset included traffic, infrastructural, and environmental variables that potentially affect speed dispersion. The Random Forest results revealed that the percentage of speed limit violations was the most influential predictor, showing a strong positive association with the magnitude of speed variability. Average speed exhibited a nonlinear effect, remaining relatively stable at lower levels but sharply amplifying variance at higher speeds. Moreover, a moderate share of heavy vehicles (30–40%) increased speed fluctuations, whereas higher proportions contributed to more uniform traffic flow. Among infrastructural attributes, the number of lanes had a positive impact on speed variance, reflecting the greater freedom for lane changes and maneuvering, while the presence of speed bumps significantly reduced speed dispersion. Environmentally, higher roadside population density and agricultural land use were associated with decreased variance, suggesting more cautious driving behavior in such areas.

Overall, the findings highlight the necessity of integrating traffic enforcement, infrastructure design, and environmental planning into comprehensive speed management strategies. The proposed machine learning framework provides a data-driven foundation for identifying high-risk segments and informing targeted interventions to enhance rural road safety.
Keywords
Speed variance
Road safety
Random Forest
Machine learning
Traffic behavior
Subjects
-سازمان پزشکی قانونی کشور. (1402). آمار متوفیات و مصدومین حوادث رانندگی در سال ۱۴۰۲. تهران، سازمان پزشکی قانونی کشور.
-Alizadeh, M., Davoodi, S. R., & Shaaban, K. (2023). Drivers’ speeding behavior in residential streets: A structural equation modeling approach. Infrastructures, 8(1), 11-12.
-Breiman, L. (2001). Random forests. Machine Learning, 45(1), 5–32.
-Hu, W. (2017). Raising the speed limit from 75 to 80 mph on Utah rural interstates: Effects on vehicle speeds and speed variance. Journal of Safety Research, 61, 83–92.
-Mohanty, M., & Dey, P. P. (2019). Major stream delay under limited priority conditions. Journal of Transportation Engineering, Part A: Systems, 145(3), 1–10.
-Yan, M., & Shen, Y. (2022). Traffic accident severity prediction based on random forest. Sustainability, 14(3), 1729.
-Zhong, L., & Hou, D. (2010, August). Evaluation of work zone speed reduction measures. In ICCTP 2010: Integrated Transportation Systems – Green, Intelligent, Reliable, Beijing: ASCE.267–275.
-Alomari, A. H., Al-Salem, M.,Al-Basyouny, K., & El-Basyouny, S. (2023). Modeling speed mean and variance for different enforcement conditions on multilane highways. Journal of Transportation Engineering, Part A: Systems, 149(8), 04023067.
-Garber, N. J., & Gadirau, R. (1988). Speed variance and its influence on accidents. Washington, DC: Transportation Research Record.
-Gargoum, S. A., & El-Basyouny, K. (2016). Exploring the association between speed and safety: A path analysis approach. Accident Analysis & Prevention, 93, 32–40.
-Graves, P. E., Lee, D. R., & Sexton, R. L. (1993). Speed variance, enforcement, and the optimal speed limit. Economics Letters, 42(2–3), 237–243.
-Huang, B., Zhao, L., Wang, W., & Liu, H. (2013). A simulation study for minimizing operating speed variation of multilane highways by controlling access.Procedia – Social and Behavioral Sciences, 96, 2767–2781.
-Lord, D., Qin, X., & Geedipally, S. R. (2021). Highway safety analytics and modeling. Elsevier.
-Panschin, S., & Vitikainen, A. (2010, April). Land consolidation and agricultural traffic: Instrument for calculating the traffic volume. In XXIV FIG International Congress, Sydney, Australia (pp. 11–16). FIG – International Federation of Surveyors.
-Quddus, M. (2013). Exploring the relationship between average speed, speed variation, and accident rates using spatial statistical models and GIS. Journal of Transportation Safety & Security, 5(1),
27–45.
-Roh, C. G., Jeon, H., & Son, B. (2021). Do heavy vehicles always have a negative effect on traffic flow? Applied Sciences, 11(12), 1–15.
-Shen, J., & Yang, G. (2020). Crash risk assessment for heterogeneous traffic and different vehicle-following patterns using microscopic traffic flow data. Sustainability, 12(23), 1–18.
  -Wang, X., Abdel-Aty, M., & Lee, J. (2018). Speed, speed variation, and crash relationships for urban arterials. Accident Analysis & Prevention, 113, 236–243.
-World Health Organization. (2023). Global status report on road safety 2023. Geneva: World Health Organization.
-Xiaobao, Y. Y. (2008). Changes in traffic characteristics affected by number of lanes on freeways. Beijing: China Academy of Transportation Research.