Prediction of Road Accidents’ Severity on Russian Roads Using Machine Learning Techniques

  • D. Donchenko
  • N. Sadovnikova
  • D. ParyginEmail author
Conference paper
Part of the Lecture Notes in Mechanical Engineering book series (LNME)


A system of road transport infrastructure is one of the key components of ensuring a population life and a normal functioning of production processes, which consist of geographically distributed interactions. Road traffic accidents’ statistics in Russia shows that the problem of road safety management remains very crucial. The use of big data and machine learning approaches is effective in developing traffic accident prediction models. Such models can significantly reduce the number of accidents according to the international experience of road safety management. The paper analyzes the possibility for the development of the road traffic accidents’ prediction model using the data provided by local police in Russia. An example of using the collected data for the development of road accident severity prediction model and analyzing which features have a huge impact on the accident severity has been provided.


Road accidents Road safety Random forest Decision trees XGBoost 



The reported study was funded by Russian Foundation for Basic Research (RFBR) according to the research project No. 18-37-20066_mol_a_ved, and by RFBR and the Government of the Volgograd region of the Russian Federation grant No. 18-47-340012_r_a. The authors express gratitude to colleagues from UCLab involved in the development of project.


  1. 1.
    Road Safety Indicators (2018) GIBDD. Accessed 11 Nov 2018
  2. 2.
    Global status report on road safety 2018 (2018) WHO. Accessed 22 Dec 2018
  3. 3.
    Statistics of accidents in Russia and the world. Dossier (2016) TASS. Accessed 10 Dec 2018
  4. 4.
    Buranov I (2017) Traffic police looking for an emergency exit. In: Kommersant. Accessed 14 Dec 2018
  5. 5.
    Sadovnikova N, Parygin D, Kalinkina M et al (2015) Models and methods for the urban transit system research. CCIS 535:488–499. Scholar
  6. 6.
    Parygin D, Sadovnikova N, Kravets A et al (2015) Cognitive and ontological modeling for decision support in the tasks of the urban transportation system development management. In: Proceedings of the sixth international IEEE conference on information, intelligence, systems and applications, Corfu, Greece, 6–8 July 2015.
  7. 7.
    Parygin D, Sadovnikova N, Kalinkina M et al (2016) Visualization of data about events in the urban environment for the decision support of the city services actions coordination. In: Proceedings of the 5th international conference on system modeling and advancement in research trends, Moradabad, India, 25–27 Nov 2016.
  8. 8.
  9. 9.
    Peltola H (2009) Evaluating road safety and safety effects using Empirical Bayesian method. Accessed 28 Oct 2018
  10. 10.
    Nambuusi BB, Brijs T, Hermans E (2014) A review of accident prediction models for road intersections. Accessed 15 Nov 2018
  11. 11.
    Analysis of the causes and consequences of road accidents (2018) Statsoft. Accessed 17 Nov 2018
  12. 12.
    Banushkina ON, Pechatnova EV (2015) Improving the efficiency of forecasting accidents on roads outside settlements based on the development of an expert system. In: Izvestiya AltGU. Accessed 12 Oct 2018
  13. 13.
    Yandex has developed a system for predicting traffic jams and accidents (2015) Yandex. Accessed 21 Sept 2018
  14. 14.
    Polyakov A (2017) Traffic science. In: Internet portal of the Rossiyskaya Gazeta. Accessed 30 Sept 2018
  15. 15.
    Golubev A, Chechetkin I, Parygin D et al (2016) Geospatial data generation and preprocessing tools for urban computing system development. Procedia Comput Sci 101:217–226CrossRefGoogle Scholar
  16. 16.
    Python Data Analysis Library (2018) Pydata. Accessed 3 Oct 2018
  17. 17.
    Scikit-learn (2018) Machine learning in python. Accessed 14 Oct 2018
  18. 18.
    Imbalanced-learn (2017) Welcome to imbalanced-learn documentation. Accessed 17 Oct 2018
  19. 19.
    Haixiang G, Yijing L, Shang J et al (2017) Learning from class-imbalanced data: Review of methods and applications. Expert Syst Appl 73:220–239CrossRefGoogle Scholar
  20. 20.
    XGBoost (2016) XGBoost developers. Accessed 26 Oct 2018
  21. 21.
    Brownlee J (2016) A gentle introduction to XGBoost for applied machine learning. In: Machine learning mastery. Accessed 10 Oct 2018
  22. 22.
    Classification: Precision and Recall (2018) Google. Accessed 15 Oct 2018
  23. 23.
    Precision-recall (2018) Scikit-learn. Accessed 22 Oct 2018
  24. 24.
    Plainis S, Murray IJ, Pallikaris IG (2006) Road traffic casualties: understanding the night-time death toll. Injury Prevent 12(2):125–128CrossRefGoogle Scholar
  25. 25.
    Garrido R, Bastos A, Almeida A et al (2014) Prediction of road accident severity using the ordered probit model. Transport Res Procedia 3:214–223CrossRefGoogle Scholar
  26. 26.
    Massie DL, Campbell KL (1993) Analysis of accident rates by age, gender, and time of day based on the 1990 nationwide personal transportation survey.;analysis. Accessed 7 Nov 2018
  27. 27.
    Understand your dataset with Xgboost (2018) R-Project. Accessed 5 Nov 2018
  28. 28.
    Parygin DS, Aleshkevich AA, Golubev AV et al (2018) Map data-driven assessment of urban areas accessibility. J Phys Conf Series 1015:042048Google Scholar

Copyright information

© Springer Nature Switzerland AG 2020

Authors and Affiliations

  1. 1.Volgograd State Technical UniversityVolgogradRussia

Personalised recommendations