Advancing Economic Forecasting with AI: Outperforming Traditional Econometric Approaches

This review evaluates the performance of artificial intelligence (AI)-driven predictive models compared to traditional econometric methods in economic forecasting. Through a systematic literature review of thirteen studies, this article examines applications in bond yield forecasting, market trends, retail sales, GDP, inflation, finance, trade flows, and epidemiological forecasting (e.g., COVID-۱۹). AI methods, including deep neural networks, Long Short-Term Memory (LSTM), and machine learning ensembles, are compared to econometric models such as ARIMA, Vector Autoregression (VAR), and GARCH. Findings indicate that AI models consistently outperform traditional methods, reducing forecast errors by ۲۰-۷۰% across metrics like Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and Mean Absolute Percentage Error (MAPE). Notable improvements include ۷% lower errors in bond yield forecasting and ۲۰% reduced MAPE in finance forecasting. However, challenges such as high computational requirements, data quality issues, and limited model interpretability persist. This study synthesizes these findings to assess AI's potential in enhancing economic forecasting accuracy while highlighting practical limitations for implementation.

کلیدواژه ها:

Economic forecasting ، Artificial intelligence ، Machine learning ، Deep learning ، Econometric models