Tibor Sipos

Accurate prediction of road traffic crash severity is essential for developing data-driven safety strategies and optimizing resource allocation. This study presents a predictive modeling framework that utilizes Random Forest (RF), Gradient Boosting (GB), and K-Nearest Neighbors (KNN) to estimate segment-level frequencies of fatalities, serious injuries, and slight injuries on Hungarian roadways. The model integrates an extensive array of predictor variables, including roadway geometric design features, traffic volumes, and traffic composition metrics. To address class imbalance, each severity class was modeled using resampled datasets generated via the Synthetic Minority Over-sampling Technique (SMOTE), and model performance was optimized through grid-search cross-validation for hyperparameter optimization. For the prediction of serious- and slight-injury crash counts, the Random Forest (RF) ensemble model demonstrated the most robust performance, consistently attaining test accuracies above 0.91 and coefficient of determination (R²) values exceeding 0.95. In contrast, for fatalities count prediction, the Gradient Boosting (GB) model achieved the highest accuracy (0.95), with an R² value greater than 0.87. Feature importance analysis revealed that heavy vehicle flows consistently dominate crash severity prediction. Horizontal alignment features primarily influenced fatal crashes, while capacity utilization was more relevant for slight and serious injuries, reflecting the roles of geometric design and operational conditions in shaping crash occurrence and severity. The proposed framework demonstrates the effectiveness of machine learning approaches in capturing non-linear relationships within transportation safety data and offers a scalable, interpretable tool to support evidence-based decision-making for targeted safety interventions.

The safety of vulnerable road users is a key priority for road safety worldwide. One of the most vulnerable road users group on the road is pedestrians. Every user on the road is also a pedestrian at some point in the transport chain (vehicle access; transfer between vehicles). In everyday transport, designated pedestrian crossing point (vehicle-pedestrian interaction) is one of the road safety risk points for pedestrians. Previous research has shown that several factors can influence pedestrian priority, but most of this research has looked at the problem from the pedestrian’s perspective. Our research aimed to investigate which factors increase the drivers’ willingness to give priority to pedestrians the most. We developed a self-reported questionnaire for drivers. The questionnaire included questions on the socio-demographic characteristics and driving habits of car divers and infrastructure design. The results were analyzed using descriptive statistics. We identified two factors that increase the willingness to yield to pedestrians at signalized pedestrian crossings, both collectively and at the subgroup level. Our results can be used in several areas: infrastructure planning, transport education, and enforcement.

This literature review categorizes machine learning studies in traffic accident severity prediction, providing a comprehensive overview of the diverse applications and advancements in this field. It begins with a comparative analysis of machine learning models, highlighting the performance of various algorithms such as Random Forest, XGBoost, and Support Vector Machines (SVM) in predicting accident severity. The review also explores factor-specific studies, emphasizing the influence of road, environmental, and vehicle-related factors on crash outcomes. These studies demonstrate the critical role of factors such as road type, weather conditions, and vehicle characteristics in determining accident severity. Additionally, crash-type-specific prediction models have been developed, showcasing the ability of machine learning models to tailor predictions based on the nature of the crash, whether involving pedestrians, vehicles, or specific collision types. The review also examines hybrid and ensemble approaches, which combine multiple algorithms to enhance prediction accuracy. These approaches leverage the strengths of individual models to improve overall performance, offering a promising direction for future research. By categorizing the studies into these key areas, this review provides a structured understanding of the state-of-the-art in machine learning applications for traffic accident severity prediction and identifies opportunities for further development to enhance prediction robustness, accuracy, and applicability in real-world traffic safety management.

Összefoglalás. A közlekedési balesetekben sérült személyek, valamint a baleset miatti anyagi károk jelentős mértékű társadalmi-gazdasági veszteségértéket képviselnek. Kutatásommal célom a közlekedés biztonságosabbá tételéhez szükséges kulcsterületek feltárása és a lehetséges potenciális beavatkozási területek meghatározása volt. Munkám során a 2016–2020 közötti közlekedési balesetek adatait elemeztem leíró statisztikai módszerrel. Megállapítottam, hogy a leginkább kritikus pont a gyalogosok közlekedése. Azokban a balesetekben, ahol volt gyalogos sérült, a balesetek két okcsoportja dominál: gyalogosok hibája (34,3%), valamint a járművezetők gyalogosok részére történő elsőbbségadásának elmaradása (38,1%). Az eredmények lehetőséget adnak a további célirányos vizsgálatokra, valamint a hatékony beavatkozásokra. Summary. Transport is an essential part of everybody lives. The aim of transport users is to make their journeys without accidents in the shortest possible time, while optimising their individual resources and inputs. The number of people injured in road traffic accidents with personal injury and the material damage caused by a significant socio-economic loss. By preventing accidents, it is clear that the costs and resources spent on road accidents can be reduced. Therefore, I started my research by determining the share of road user groups in the number of accidents, the number of people injured and the amount of economic loss value associated with accidents. I analysed road accidents between 2016 and 2020. By the accident analysis, I found that among vulnerable road users, the proportion of pedestrians and cyclists injured was extremely high. Injuries in these two transport modes accounted for more than half of the injuries in cars. By calculating the loss value, I found that the damage to the national economy caused by pedestrian injuries was significant. In my research, by analysing the causes of accidents, I found that in accidents where there was a pedestrian injured, two groups of causes dominate: pedestrian error (34.3%) and drivers not giving priority to pedestrians (38.1%). Examining the accident data, it was evident that the number of pedestrian accidents due to pedestrian error is very high, so I continued my research in the area of other activities of pedestrians while walking. I measured pedestrians in 10 locations in Budapest. During the measurements, I recorded and analysed data of 11536 persons. The average rate of distraction was 54.3%, which means that one in two pedestrians is doing some other activity while walking. The most frequent activity was talking to another pedestrian (29.8%), but phone use (15.7%) and head/earphone use (13.5%) were also very significant. My research results so far show that if we want to improve pedestrian safety, we need to reduce accidents caused by pedestrians’ own fault and increase drivers’ willingness to give priority to pedestrians. My research plans also include investigating ways to reduce pedestrian errors, increasing pedestrian awareness and examining pedestrian-driver interactions in more depth. I already have results in this area, but my aim is to identify target areas (including human, infrastructure, vehicle field) for my research that can be implemented in practice, thus helping to improve pedestrian safety and road safety.

The level of greenhouse gas emissions is one of the most important issues today, both professionally and politically, because a lower level of greenhouse gas emission is mandatory for a sustainable economy. Besides industry and households, the transport sector is also responsible for these emissions. For this reason, it may be essential to set up a model with which the amount of CO2 emissions could be estimated or predicted. This article presents a model that examines the extent of economic development and CO2 emissions in European countries. The result is establishing a pattern requiring a longer time series. If the pattern is proven, a clear reassessment of the current relationship between economic development and environmental protection should be made.

This research aimed to examine the spatial distribution of road traffic accidents in Budapest, Hungary. The primary objective was to identify the factors associated with traffic accidents on the city's transportation network and to determine the locations of the most frequent accidents during peak and off-peak hours. A quantitative methodology was employed in this study, utilizing a dataset of recent accidents that occurred between 2019 and 2021, classified into peak and off-peak incidents. The data was analyzed using Python software and Quantum Geographic Information System (QGIS) tools for big data analytics. These programs enabled the creation of spatial maps of the study area and the identification of accident spots based on latitude and longitude information. A decision tree classification approach was used in the machine-learning method implemented with Python software. Additionally, the dataset file was uploaded to QGIS, which applied the heatmap (Kernel Density Estimation) algorithm to identify accident hotspots. The study findings revealed that the city center was the most common location for accidents overall, with peak and off-peak times, lanes, and days of the week investigated as potential contributing factors. The target variable was the number of accidents involving serious and minor injuries, which were found to be significantly associated with the identified accidents in this study.

The relationship between economic growth and transport sector is an important and popular topic for researchers, but it also has several untapped areas. To ensure continuous economic growth, it is necessary to answer how and to what extent economic sectors contribute to sustainability; what factors or sets of factors can determine freight performance in a country or region; and how it affects the global economy. This study aims to test the presence of spatial dependence. In this research, the authors looked for the spatial relationships between economic activity (GDP) and freight transport performance using spatial econometric models. The results showed that the spatial impact of freight transport performance and GDP significantly influence each other. The intensity calculation shows that the Baltic States have a high intensity in road freight transport, followed by the Central European region. Eastern Europe, including Russia and the Baltics, are prominent players in rail freight. Furthermore, the spatial econometric models have highlighted that a country with high GDP has some sort of "suction" effect on neighbouring countries with lower GDP along with the freight performance. This is especially true for rail freight. In the long run, the outlined results may even support strategic decision-makers in managing the economic impacts of both road and rail freight transport at the regional level.