Construction and Optimization of Macroeconomic Data Forecasting Model Based on Machine Learning

Abstract

Human capital was first viewed as a production component in macroeconomic development, but endogenous growth theories eventually replaced this perspective. The majority of earlier research used econometric models to investigate the GDP forecasting. Since machine learning models can efficiently resolve nonlinear interactions, this study offers a new perspective by examining the linkages using machine learning approaches. The prediction model for economic development was created for this reason using the best machine learning techniques, specifically the Support Vector Machine. In order to improve SVM prediction, the hyper parameters were optimised using the Bayes approach and several kernel functions. Three statistical metrics—the coefficient of determination, mean absolute error, and root mean square error are used to assess the models' effectiveness.