INVESTIGATION OF THE ACCURACY OF THE METHODOLOGY FOR PREDICTING SOURCES OF NATURAL EMERGENCIES USING ARTIFICIAL NEURAL NETWORKS

An analysis of modern databases on registered natural and man-made emergency situations shows their rapid increase, despite the complex of technical and organizational measures taken. The work presents the main factors contributing to the development of emergency situations The current issue of the need to improve the early warning system for natural disasters is considered. The authors of the work presented an original computer program that implements the operation of artificial neural networks capable of predicting the likelihood of the occurrence of hazardous natural meteorological phenomena. The results obtained showed satisfactory accuracy, as evidenced by regression analysis. The practical significance of the study lies in the ability to convert observation data into practical actions to prevent emergency situations of various types.