Work

Improving the Methodology for Determining Injury Rates in Road Accidents using Machine Learning - The Random Tree Method

Year: 2025

Type: conference-paper

Abstract: Every year in Russia there are a large number of road traffic accidents (RTA), in which more than 100 thousand people are injured and over 10 thousand die. (https://rg.ru/2023/11/18/ezhegodno-v-avtomobilnyhavariiah-gibnet-bolee-milliona-chelovek-chto-mozhet-izmenit-pechalnuiu-statistiku.html, 2023). This has an impact on the demographic and economic situation of the country. The aggravated problem forces the Government to develop various programs and strategies to improve road safety. They are primarily aimed at preventing accidents with the help of sanctions policy. But the practice of holding drivers accountable is not always objective and leads away from the truth of the origin of accidents, especially with victims (Pugachev, I., Furman, B., Umanets, I., 2023). The article proposes a comprehensive approach to identifying accident blackspots with a high risk of injury. The study includes an analysis of the dependencies of all elements of the Driver-Vehicle-Road-Environment (DVRE) system. Regression analysis and machine learning libraries of the random forest method in the Python programming language are used.

DOI: 10.63550/iceip.2025.1.1.032

Authors: I.N. Pugachev, D.Y. Grigorov, P.B. Skripko, N.G. Sheshera

Institutions: Far Eastern Branch of the Russian Academy of Sciences, Ministry of Internal Affairs of the Russian Federation, Far Eastern Law Institute of Ministry of Internal Affairs of Russia

Language: English

Topic: Traffic and Road Safety

Subfield: Safety, Risk, Reliability and Quality

Field: Engineering

Domain: Physical Sciences

SDGs: Good health and well-being, Sustainable cities and communities, Industry, innovation and infrastructure

Open Access status: gold

OpenAlex ID: W4410317031

FWCI:

Cites:

Cited by:

Related to: 10