Sheikhmohammadi Amir, Hosseinpour Saeed, Jalilzadeh Zahra, Azarpira Hossein, Yousefi Mahmood
Department of Environmental Health Engineering, School of Health, Khoy University of Medical Sciences, Khoy, Iran.
Department of Environmental Health Engineering, School of Public Health, Urmia University of Medical Sciences, Urmia, Iran.
Sci Rep. 2025 May 29;15(1):18852. doi: 10.1038/s41598-025-03814-x.
This paper assesses the presentation of Gradient Boosting Regression (GBR), Ridge Regression (RR), and Particle Swarm Optimization (PSO) models in improving the photocatalytic destruction of antibiotic utilizing a UV/ZrO₂/NaOCl system. The GBR model indicated the strength of exhibited precision, with high R² and Explained Variance Score (EVS) esteems, however gave indications of overfitting. Remarkably, the RR model obtained had many significant values and the model had a high data fitness with an R² of 0.81; however, error bars were observed in some areas that could be optimized by fine-tuning. Feature significance examination indicated that X2, X1, and X5 fundamentally affected the model's performance. At last, PSO was instrumental in finding the ideal limit mix to support removal performance. Altogether the presented models help to improve the photocatalytic degradation process and provide useful tools for further investigation in this field.
本文评估了梯度提升回归（GBR）、岭回归（RR）和粒子群优化（PSO）模型在利用UV/ZrO₂/NaOCl系统改善抗生素光催化降解方面的表现。GBR模型显示出较高的精度，具有较高的R²和解释方差得分（EVS）值，但有过拟合迹象。值得注意的是，所获得的RR模型有许多显著值，且该模型的数据拟合度较高，R²为0.81；然而，在某些区域观察到误差条，可以通过微调进行优化。特征重要性检验表明，X2、X1和X5对模型性能有根本性影响。最后，PSO有助于找到理想的极限组合以支持去除性能。总体而言，所提出的模型有助于改善光催化降解过程，并为该领域的进一步研究提供有用工具。
Zotero 插件
