Forecasting the Number of Road Accidents in Poland and Hungary Using Neural Networks

Abstract

The incidence of road accidents in Poland and Hungary has been on a downward trend annually, a pattern that is also evident globally. Although the recent pandemic has impacted these statistics, the total remains alarmingly high. Consequently, it is crucial to implement all possible strategies to reduce this figure further. This article aims to project future road accidents in Poland and Hungary. To achieve this, we analysed yearly statistics regarding road accidents in the two nations. Utilising data from the Polish Police and the Hungarian Central Statistical Office, we made forecasts from 2024 to 2030. Several selected neural network models were employed for this predictive analysis. The findings indicate that we can anticipate a continued stabilisation in the number of road accidents. Several factors, including the rising number of vehicles on the roads and the development of new highways, influence this trend. Additionally, the chosen sizes for the sample sets (learning, testing, and validation) play a significant role in the results obtained.