Comparative Analysis of Machine Learning and Traditional Models in Economic Forecasting

Abstract

Economic forecasting is pivotal for decision-making across governmental, business, and individual domains, driving sustainable development and resource management amid growing global economic uncertainties. This study examines the application of both machine learning and traditional models in economic forecasting, highlighting their respective strengths and limitations. Machine learning models, especially deep learning, demonstrate superior performance in handling large-scale, high-dimensional, and nonlinear relationships, enhancing predictive accuracy. Traditional models, such as linear regression and time series, maintain advantages in interpretability and uncertainty modeling, crucial for small samples and scenarios demanding transparent decision-making. By comparing these models in terms of performance, interpretability, and uncertainty management, the research underscores the complementary nature of machine learning and traditional approaches. The findings suggest that the integration of these methodologies can offer more robust and flexible solutions for economic forecasting, addressing complex and dynamic economic environments. Future research should focus on developing hyb