Classification Trees in the Assessment of the Road–Railway Accidents Mortality

Abstract

A special element of road safety research is accidents at the interface of the road and rail system. Due to their low share in the total number of incidents, they are not a popular subject of analyses but rather an element of collective studies, whereas the specificity of the road–rail accidents requires a separate characteristic, allowing, on the one hand, to categorize these types of incidents, and on the other, to specify the factors that affect them, along with an assessment of the strength of this impact. It is important to include in such analyses all potential predictors, both qualitative and quantitative. Moreover, the literature considers most often a number of accidents while, according to the authors, it does not fully reflect the scale of the danger. A better evaluation would be the victim’s degree of injury. Therefore, the purpose of this article is to assess the likelihood of occurrence of various effects of road–rail accidents in the aspect of selected factors. Due to the ordinal form of the dependent variable, the classification trees method was used. The results obtained not only allow the characterization and assessment of the danger but also constitute guidelines for taking preventive actions.

Construction of Regression Models Predicting Lead Times and Classification Models

“…The predictors have the same characteristics as for regression. The XGBoost (Kozłowski 2021) classifier was used for classification. The model was also built on data from the last two years (2020-2021).…”

Minimizing Cost Overrun in Rail Projects through 5D-BIM: A Systematic Literature Review

“…Although the extant literature on the topic is limited, the findings suggest an emerging tendency for artificial intelligence (AI) and machine learning (ML) publications in the railway domain [290]. Currently, rail projects are utilizing AI and ML to optimize the effectiveness and performance of railway systems by different means, including resources and equipment planning during the construction stage [291], delving into the causation factors of highway-rail crossing crashes [292], categorizing fatality rates for accidents [293], improving safety measures [294,295], mitigating collision risks [296], and integrating building information modeling (BIM) and ML to enhance the operation and maintenance of railway networks [297].…”

Data-Driven Analysis of Fatal Urban Traffic Accident Characteristics and Safety Enhancement Research

“…Oralhan et al [18] used cox regression and life table model to analyze the traffic accident data in Kayseri, Turkey, and obtained the influencing factors of fatal accidents. Kozlowski et al [19] used an improved classification tree algorithm to classify and evaluate accident factors in road-rail intersections in the Polish region to qualify and assess them.…”