FORECASTING METHODS FOR STUDYING AND DETECTING NATURAL PHENOMENA

Authors

DOI:

https://doi.org/10.32782/IT/2023-4-2

Keywords:

natural phenomena, disaster prediction, perception architecture, neural network, prediction method, disaster monitoring.

Abstract

This article describes ways to research and detect natural phenomena based on forecasting methods and techniques. The purpose of this work is creation of improved algorithm that will allow predicting the occurrence of any kind of natural phenomena based on existing statistics. To create proposed algorithm and software, we used existing forecasting methods and techniques, mathematical and causal methods as well as monitoring current affairs will be considered. Proposed algorithm improvements give us possibility to get general prediction or get prediction for some specific kinds of disaster. Also in that article we propose to combine mathematical methods together with artificial intelligence. AI allow us improve accuracy of prediction and provide possibility to increase number of parameters or characteristics to analysis. As AI is modern and fast-growing technology it provides unlimited ways to improve our algorithm and software not only for forecasting of natural phenomena but also for simulate them, analyze consequences, ways to minimize damage and most important – casualties. One of the main advantages of using proposed combining artificial intelligence with old mathematical and statistics methods over using only mathematical or statistical methods is a flexibility of artificial intelligence in their result as mathematical result stabilize with growing statistics data and each new occurrence will not take so big impact on result. But for artificial intelligence each new data can have a critical effect and can correct all forecasts together with expected consequences. As result of that article new software complex will be implemented and integrated to scientific complex for further improvements, learnings, researches and analysis.

References

Бурков А. Машинне навчання без зайвих слів. Print2print. 2019.

Горник К., Стінчкомб М., Уайт Г. Багаторівневі мережі прямого зв'язку є універсальними апроксиматорами. Neural Networks. 1999.

Тейлор С. Д. Наука про стихійні лиха: Коли природа і люди стикаються. Nomad Press. 2020.

Global Disaster Alert and Coordination System : веб-сайт. URL: https://www.gdacs.org (дата звернення: 10.09.2023).

Кіпроно Е.К. Як працює зворотне поширення в нейронних мережах? 2022. Towards Data

Science : веб-сайт. URL: https://towardsdatascience.com/how-does-back-propagation-work-in-neuralnetworks-with-worked-example-bc59dfb97f48 (дата звернення: 12.09.2023).

Джозеф Р. Зміна клімату: що потрібно знати кожному. Oxford University Press. 2015. 328 с.

Чару К. А. Нейронні мережі та глибоке навчання: Підручник. Springer. 2018. 520 c.

Льюїс Н.Д. Нейронні мережі для прогнозування часових рядів за допомогою R. Apress. 2018. 700 с.

ECMWF : веб-сайт. URL: https://www.ecmwf.int/ (дата звернення: 12.09.2023).

МакГілліврей П. Передбачення непередбачуваного : веб-сайт. URL: https://www.uvic.ca/knowledge/archives/2010/v10n04-april10.php (дата звернення: 12.09.2023).

Downloads

Published

2023-12-28