Short-Term Economic Forecasting by Complex-Valued Autoregressions

Abstract

One of the directions that can expand the instrumental base for modeling the economy is complex-valued economics – ​a section of economic and mathematical modeling devoted to the use of models and methods of the theory of the function of a complex variable in economics. The article discusses the possibility of short-term economic forecasting using autoregressive models of complex variables. A classification of possible modifications of complex-valued autoregressive models is given, and the main properties of each of the classes of these models are shown. One of the varieties of these complex-valued models uses current and past errors of approximation, which means that it can be compared with the widely used model of autoregressive real variables ARIMA(p, d, q). The article makes such a comparison, both on a theoretical level and on a practical example.