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

V. I. Baluta


Abstract: This work is devoted to the review and analysis of information materials on the problem of synthetic data formation in the preparation of training sets of predictive and analytical models in emergency response systems.
A brief 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 various forms 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 of threats based on expert ratings.

Keywords: artificial intelligence, synthetic data, data generation, machine learning, mathematical modeling.

DOI: 10.20948/prepr-2022-60