Algwil Ahmed Ramdan Almaqtouf, Khalifa Wagdi M S
Cyprus Health and Social Sciences University, Mersin 10, Turkey.
University of Mediterranean Karpasia, Mersin 10, Turkey.
Sci Rep. 2025 Apr 8;15(1):11948. doi: 10.1038/s41598-025-95678-4.
This study introduces a novel hybrid model for accurate CO emissions prediction, supporting sustainable decision-making. The model integrates the Gaussian mutation and shrink mechanism-based moth flame optimization (GMSMFO) algorithm with an extreme learning machine (ELM). GMSMFO enhances population diversity and avoids local optima through Gaussian mutation (GM), while the shrink mechanism (SM) improves exploration-exploitation balance. Validated on the congress on evolutionary computation (CEC2020) benchmark suite (dimensions 30 and 50), GMSMFO demonstrated superior performance compared to other optimization algorithms. Applied to fine-tune ELM parameters, the GMSMFO-ELM model achieved exceptional predictive accuracy, with a coefficient of determination (R) of 96.5%, outperforming other hybrid models across metrics such as root mean squared error (RMSE), normalized root mean squared error (NRMSE), mean absolute error (MAE), and mean square error (MSE). Feature importance analysis highlighted economic growth, foreign direct investment, and renewable energy as key predictors. This study highlights the robustness and adaptability of GMSMFO-ELM, establishing it as a reliable framework for advancing global sustainability objectives.
本研究引入了一种用于准确预测一氧化碳排放的新型混合模型，以支持可持续决策。该模型将基于高斯变异和收缩机制的蛾火焰优化（GMSMFO）算法与极限学习机（ELM）相结合。GMSMFO通过高斯变异（GM）增强种群多样性并避免局部最优，而收缩机制（SM）改善探索-利用平衡。在进化计算大会（CEC2020）基准测试套件（维度为30和50）上进行验证，GMSMFO与其他优化算法相比表现出卓越性能。应用于微调ELM参数时，GMSMFO-ELM模型实现了出色的预测精度，判定系数（R）为96.5%，在均方根误差（RMSE）、归一化均方根误差（NRMSE）、平均绝对误差（MAE）和均方误差（MSE）等指标上优于其他混合模型。特征重要性分析突出了经济增长、外国直接投资和可再生能源作为关键预测因素。本研究突出了GMSMFO-ELM的稳健性和适应性，将其确立为推进全球可持续发展目标的可靠框架。
