Improving the accuracy of economic growth rate forecasting using a combination of wavelet transform and Artificial Neural Network

Articles in Press

Document Type : Original Article

Authors

1 Ph.D. in Economics (Public Sector Economics), Auditor of the Kurdistan Province Tax Office, Sanandaj, Iran.

2 Assistant Professor, Department of Economics, Payam Noor University, Tehran, Iran.

Abstract

Forecasting the behavior of macroeconomic and financial variables of the economy is of great importance to policymakers. This is a challenging task for developing economies, including Iran, because a set of factors that are not considered in the main economic theories often play an important role in shaping the environment and overall outlook of their economy. Economic relations in such environments are more unstable and accompanied by nonlinear trends. The combination of wavelet transform and artificial neural network is an emerging mathematical modeling analysis in recent years that has been proposed to denoise time series and increase forecast accuracy. In this study, an attempt was made to present a model for predicting Iran's economic growth using the combination of wavelet transform and artificial neural network to compare the forecast accuracy of the combined method with the artificial neural network. The results of the study showed a significant improvement in the forecasting of the neural network using denoised data. The results also confirmed the superiority of the combined model over the XGBoost and ARIMA models. The prediction accuracy of these models has been evaluated and compared based on criteria such as Root Mean Square Error and the Diebold-Mariano test.

Keywords

Main Subjects

Economic Policies and Research

Articles in Press, Accepted Manuscript
Available Online from 09 December 2024

Files

History

  • Receive Date: 06 September 2024
  • Revise Date: 08 December 2024
  • Accept Date: 09 December 2024
  • Publish Date: 09 December 2024

Share

How to cite

Statistics