METHODS OF FORECASTING AND DATA CLASSIFICATION BASED ON NEURAL NETWORKS

Abstract

The article is devoted to a comprehensive review of neural network models in forecasting and classification tasks.Finding the strengths and weaknesses of different forecasting methods using neural networks. Exploring the possibilities of improving the use of neural networks in forecasting and classification tasks.The purpose of the work. The purpose of this work is to study methods of forecasting and data classification based on neural networks. Which means a review of existing approaches and finding new ways to improve the solution of the above problems. Finding ways to improve existing models. The task of this study is to compare existing methods of using neural networks in forecasting problems and to obtain new approaches to improve existing methods.Methodology. It is based on the analysis of scientific publications on neural network models, as well as prediction and classification methods. For this purpose, the characteristics and methods of comparative analysis of the strengths and weaknesses of neural networks are provided. As well as recommendations for improving prediction methods, where possible.Scientific novelty. The solution to the problems set and the scientific novelty of this research lies in identifying ways to improve methods and in a criterion-based comparison of existing methods for using neural networks in data classification and prediction tasks, improving new approaches based on existing ones to improve the processing processes of the above- mentioned tasks.Conclusions. Analysis of neural network models in forecasting tasks revealed their strengths and weaknesses. Criterion analysis established the advantages of forecasting methods using neural networks. Recommendations for improving forecasting methods are proposed.