Estimating Relevant Exchange Rate Fluctuations Using a Novel Multivariate Gray Model with GOWA Operator

Objective and accurate exchange rate estimations are vital for policy-formulating in the central bank and associated sectors. The data of exchange rate non-stationarity, nonlinearity and lag effects, this paper develops a novel gray multivariate generalized ordered weighted average operator model (GM-GOWA) based on the GOWA operator. In this new model, the weight vector is used to alleviate data non-stationarity and the whale optimization algorithm is employed to further reduce the influence of other factors on data stationarity, the parameter [Formula: see text] in the GOWA operator is utilized to extract the nonlinear relationship between variables, which enhances the predictive ability of the gray system, and the lag effect is also used to make full use of internal data information. The new model is compared with other 11 models, in two exchange rate cases: USD/INR, USD/ISK. Results show that the GM-GOWA model is superior to competitive models, proving its stronger flexibility and applicability.