Информационно-аналитическая модель прогнозирования возникновения опасных метеоусловий для оповещения населения

Abstract

Introduction. Today, security issues in society are more relevant than ever. One of the most important components of safety is the control of weather conditions and monitoring of weather phenomena. Prompt and accurate information about expected meteorological conditions can prevent catastrophic consequences for the population and infrastructure. Therefore, forecasting meteorological phenomena is extremely important for society and the state. The objective of the study. Increasing the level of readiness of the public warning system by developing tools to ensure the safety of people based on the systematization of means of identifying sources of emergencies (including using space technologies), forecasting their development in time and space, as well as a public warning system. Methods. The model is a set of methods of mathematical statistics and machine learning that allow you to analyze the collected data and build forecasts of possible dangerous meteorological phenomena based on them. This paper uses the method of regression analysis, the retrospective method, the Bayesian approach, as well as system analysis and synthesis. Results and discussion. As a result of the study, an assessment of the dynamics of notification before the onset of emergency hazards was carried out. A prognostic model of the hazard level based on multivariate regression analysis has been developed, which includes modeling the threats of dangerous meteorological phenomena using the Bayesian approach. The effectiveness of the developed information-analytical model for predicting the occurrence of dangerous weather conditions for alerting the population is determined. The developed methods are implemented in the computer program "Determination of the level of danger caused by dangerous meteorological phenomena" on the example of the Republic of Crimea. Conclusions. An information and analytical model for predicting the occurrence of dangerous meteorological conditions to alert the population is an important task of the entire state security system. The development of methods of mathematical statistics and machine learning makes it possible to create effective forecasting models that provide timely notification of the population about possible dangerous meteorological phenomena. Keywords: information-analytical model; warning of the population; dangerous meteorological conditions; machine learning; forecasting.