Predicting Financial Outcomes from Environmental Costs in Shariah Green Accounting

Abstract

This study aims to explore the relationship between environmental costs and financial performance in the Islamic finance sector using advanced machine learning techniques. By integrating CSR Expenditure, Energy Consumption, and Carbon Emissions as environmental factors, the research applies Gradient Boosting, XGBoost, and Artificial Neural Networks (ANN) to predict Retail Banking Revenue, Wholesale Banking Revenue, and Third-Party Funds. The objective is to evaluate how sustainability practices impact financial outcomes, using an innovative approach that combines economic modeling with computer-based prediction. The findings reveal that Gradient Boosting outperforms other models, demonstrating strong predictive accuracy, especially for Third-Party Funds and Wholesale Banking Revenue. XGBoost also provides valuable insights, while ANN struggles with overestimations, indicating the need for further optimization. This research underscores the growing significance of environmental sustainability in shaping financial performance and provides a computational framework for financial institutions and policymakers to assess the impact of green accounting on economic growth.