Faculty of Civil and Environmental Engineering, Near East University, Near East Boulevard, Via Mersin, 99138, Nicosia, North Cyprus, Turkey.
Center of Excellence in Hydroinformatics, Faculty of Civil Engineering, University of Tabriz, Tabriz, Iran.
Environ Sci Pollut Res Int. 2021 Sep;28(36):49663-49677. doi: 10.1007/s11356-021-14133-9. Epub 2021 May 3.
Accuracy in the prediction of the particulate matter (PM and PM) concentration in the atmosphere is essential for both its monitoring and control. In this study, a novel neuro fuzzy ensemble (NF-E) model was proposed for prediction of hourly PM and PM concentration. The NF-E involves careful selection of relevant input parameters for base modelling and using an adaptive neuro fuzzy inference system (ANFIS) model as a nonlinear kernel for obtaining ensemble output. The four base models used include ANFIS, artificial neural network (ANN), support vector regression (SVR) and multilinear regression (MLR). The dominant input parameters for developing the base models were selected using two nonlinear approaches (mutual information and single-input single-output ANN-based sensitivity analysis) and a conventional Pearson correlation coefficient. The NF-E model was found to predict both PM and PM with higher generalization ability and least error. The NF-E model outperformed all the single base models and other linear ensemble techniques with a Nash-Sutcliffe efficiency (NSE) of 0.9594 and 0.9865, mean absolute error (MAE) of 1.63 μg/m and 1.66 μg/m and BIAS of 0.0760 and 0.0340 in the testing stage for PM and PM, respectively. The NF-E could improve the efficiency of other models by 4-22% for PM and 3-20% for PM depending on the model.
准确预测大气中颗粒物（PM 和 PM）的浓度对于监测和控制都至关重要。本研究提出了一种新颖的神经模糊集成（NF-E）模型，用于预测每小时 PM 和 PM 浓度。NF-E 涉及仔细选择相关输入参数进行基础建模，并使用自适应神经模糊推理系统（ANFIS）模型作为非线性核来获得集成输出。所使用的四个基础模型包括 ANFIS、人工神经网络（ANN）、支持向量回归（SVR）和多元线性回归（MLR）。使用两种非线性方法（互信息和单输入单输出 ANN 基于灵敏度分析）和传统的 Pearson 相关系数选择用于开发基础模型的主导输入参数。NF-E 模型被发现能够更准确地预测 PM 和 PM，具有更高的泛化能力和最小的误差。NF-E 模型优于所有单一基础模型和其他线性集成技术，在测试阶段，PM 和 PM 的纳什-苏特克利夫效率（NSE）分别为 0.9594 和 0.9865，平均绝对误差（MAE）分别为 1.63 μg/m 和 1.66 μg/m，偏差分别为 0.0760 和 0.0340。NF-E 可以根据模型的不同，将其他模型的效率提高 4-22%用于 PM 和 3-20%用于 PM。
