Sydor P. Methods of forecasting natural disasters based on artificial intelligence technologies

The object of the study is the processes of planning and organizing safe tourist trips, taking into account the risks of natural disasters. The subject of the study is methods for predicting natural disasters (forest fires, hurricanes, floods) based on artificial intelligence technologies. The aim of the work is to develop an innovative information technology for planning safe tourist trips, which is based on advanced methods of predicting natural disasters using artificial intelligence technologies. The following methods were used as research tools: linear models, ANFIS, neural networks, LSTM, correlation analysis, R/S analysis, decision trees, and ensemble models. The scientific novelty of the obtained results: 1. An MLOps technology was developed for systems with small and medium volumes of input data by implementing an imperative model. 2. Mathematical models for predicting forest fires were improved. 3. Mathematical models for predicting hurricanes were improved. 4. Classification ensembles were improved. 5. An information technology for forecasting crisis phenomena was developed. The practical significance of the obtained results: The developed methods for predicting natural disasters were implemented into information technologies, contributing to the prompt receipt of forecasts for risk reduction and disaster impact mitigation. These methods can be used by government authorities, environmental organizations, and businesses to manage natural disaster risks and help protect the population and infrastructure. Integrating these methods into mobile applications ensures widespread dissemination of information among the population. Implementation acts: 1. Department of Investment Policy and Tourism of the Chernivtsi Regional State Administration: implemented forecasting methods for analyzing the risks of natural disasters in tourist routes (act dated May 14, 2024). 2. NGO "RTO 'Hospitable Bukovyna'": developing an information system warning about disaster risks for tourist routes (act dated May 2, 2024). 3. Institute of Geography of the Serbian Academy of Sciences: improved the quality of scientific research and citations through the implementation of forecasting methods (act dated April 29, 2024).