A Sustainable Asset Portfolio Management Approach: Integrating Support Vector Regression and GARCH Model for Green Investments

Abstract

This study presents an integrated Support Vector Regression and GARCH (SVR-G) model tailored for simulating synthetic green investment asset returns, which are characterized by unique financial properties, including non-normal return distributions, higher volatility, and positive skewness. Unlike traditional assets, green investments often exhibit irregular return patterns that require advanced predictive and optimization methods. To address these complexities, synthetic data was generated using GARCH(1,1) processes and skew-t distributions to capture the asymmetry and heavy tails typical of green asset returns, thereby simulating realistic market conditions. Through hyperparameter tuning and K-Fold Cross-Validation, the model achieved high predictive accuracy, with a correlation coefficient of 0.9992, Nash-Sutcliffe Efficiency of 0.9962, and a minimal Root-Mean-Square Error of 0.0053, demonstrating its robustness and strong alignment with actual returns. Despite its strengths, the model’s reliance on synthetic data highlights the need for validation using real-world datasets and the exploration of additional risk metrics. By leveraging GARCH(1,1) processes and skew-t distributions, this research demonstrates the SVR-G model's capacity to generate realistic synthetic datasets for green assets, enabling improved analysis and decision-making for sustainable investment strategies.