Formation of synthetic data in the preparation of training sets for emergency warning systems

Abstract

The preprint contains a brief overview of information materials on the problem of synthetic data formation and substantiation of the possibility of using mathematical modeling to prepare training sets in the interests of creating predictive and analytical models used in emergency warning and response systems by means of machine learning. An assessment of the current state of research in the field of machine learning is given, an analysis and generalization of practical methods for obtaining synthetic data for the formation of training sets with detailed features in relation to the presentation of information in the form of numerical, textual or figurative formats is carried out, recommendations are made on the use of various mechanisms for creating synthetic data in order to prepare training sets of predictive and analytical models for the main types threats.