Modern methods of time series forecasting

Abstract

Purpose: the relevance of studying time series has recently reached a new level of popularity in various scientific fields and in economics, in particular, therefore, in this article we will review the concepts and components of time series, then discuss some common forecasting methods. Discussion: аccording to the described research objectives and based on the results of the analysis of the concept of a time series and methods of forecasting time series, including series with memory, the authors identified the most progressive methods, namely adaptive, neural networks and cellular automaton. Results: the result of the study is the disclosure of the key points of the application of certain forecasting methods, depending on the component composition and origin of the time series under study.