Predictive Modeling of the New Energy Index Trading Volume in Mainland China: A Gaussian Process Regression Approach in Machine Learning

The accurate prediction of energy index trading volumes remains a critical priority for market participants and regulatory bodies. This research addresses the gap in existing literature by focusing on daily trading volume prediction for China’s mainland new energy index between January 2016 and December 2020, a vital financial indicator that has received limited attention. The analytical framework employs Gaussian process regression (GPR) techniques, with model training enhanced through 10-fold cross-validation and Bayesian hyperparameter optimization. Empirical validation demonstrates the model’s effectiveness, achieving an out-of-sample relative root mean square error (RRMSE) of 17.1352% during the 2020 test period, aligning with established accuracy benchmarks in econometric forecasting. These predictive tools offer practical value for investment strategy formulation and regulatory policy development, enabling data-driven decision-making processes. Furthermore, the methodological framework demonstrates transferability potential, providing insights applicable to the design and evaluation of analogous energy indices across financial markets.