Using neural networks to forecast the frequency of accidents on particular Polish road types
Acta Tecnología · 2024
DOI: 10.22306/atec.v10i2.199 ↗Kurz gefasst
Fachartikel in der Zeitschrift Acta Tecnología, 2024. Verfasst von Piotr Gorzelanczyk und 1 weiteren Autor:innen.
Abstract (Originalsprache)
Globally, the quantity of road accidents is steadily decreasing annually. Even if the epidemic has recently had an effect, this number is still quite high. Therefore, every attempt should be made to reduce this number. The purpose of this article is to forecast the number of accidents that will transpire on various kinds of roads in Poland. For this purpose, monthly data on the total number of traffic accidents in Poland broken down by type of road were examined. Based on police data, a forecast was created for the years 2022–2040. A few neural network models were used to predict the frequency of accidents in Poland. The results show that we should keep expecting a high degree of stability in terms of the quantity of traffic accidents. On the one hand, the construction of new highways and the increase in the number of cars on Polish roads have an effect on this. The number of random samples utilized in the processes of learning, testing, and validation affects the results. In the case of 70-15-15, the average projected number of accidents for 2022-2024 should be: motorway - 827, expressway - 1136, 2 one-way carriageway - 5451, single carriageway -3067 and 1 carriageway 2 directions - 28048, respectively. For 80-10-10, these values should be respectively: motorway - 1057, expressway - 1141, 2 one-way carriageway - 5486, single carriageway -2745 and 1 carriageway 2 directions - 28039. As you can see, the values are similar.
Zitieren
APA
Gorzelanczyk, P., & Garbarova, M. (2024). Using neural networks to forecast the frequency of accidents on particular Polish road types. Acta Tecnología. https://doi.org/10.22306/atec.v10i2.199
Vancouver
Gorzelanczyk P, Garbarova M. Using neural networks to forecast the frequency of accidents on particular Polish road types. Acta Tecnología. 2024. doi:10.22306/atec.v10i2.199
BibTeX
@article{Gorzelanczyk2024,
author = {Gorzelanczyk, Piotr and Garbarova, Miriam},
title = {Using neural networks to forecast the frequency of accidents on particular Polish road types},
journal = {Acta Tecnología},
year = {2024},
doi = {10.22306/atec.v10i2.199}
}
EUVIMED