Predictive Accident Forecastin Software in Hungary and Serbia

Doi: https://doi.org/10.58603/PQUF8085

Rezumat: This paper examines predictive software tools developed in Hungary and Serbia for forecasting traffic accidents. Using artificial intelligence and statistical analysis, these systems aim to identify high-risk locations and reduce accident rates. The Serbian ANN1 and ANN2 models use neural networks to predict accident numbers and severity, while Hungary's Sopianae software integrates historical data with lunar and weather factors. These predictive technologies are particularly valuable in regions where traditional traffic safety methods have proven insufficient, and where law enforcement resources are limited. Both systems enhance preventive policing and support traffic safety planning. By analyzing temporal and spatial accident patterns, authorities can deploy patrol units more effectively and proactively reduce risks. The study evaluates the potential of predictive analytics to support the EU's long-term road safety goals. Results highlight the importance of data-driven decision-making and the need for international cooperation in preventing traffic accidents. Additionally, the paper explores the origins of predictive policing in Hungary, predating well-known American initiatives, and presents preliminary results from the Sopianae system’s application in Pécs. By promoting innovation and cross-border collaboration, these tools may significantly contribute to reducing fatalities and improving public safety, especially in Central and Eastern Europe, where predictive solutions remain underutilized.