基于MGWO-SVR的空气质量预测  

Air Quality Prediction Based on MGWO-SVR

作　　者：张楠[1] 王鹏[1] 白艳萍[1] 王永杰[1] ZHANG Nan;WANG Peng;BAI Yan-ping;WANG Yong-jie(Faculty of Science, Department of Maths, North University of China, Taiyuan 030051, China)

机构地区：[1]中北大学理学院

出　　处：《数学的实践与认识》2018年第8期159-165,共7页Mathematics in Practice and Theory

基　　金：国家自然科学基金(61774137);山西省自然科学基金(201701D22111439,201701D221121);山西省回国留学人员科研项目(2016-088)

摘　　要：空气质量指数预测可以为企业和社会工作提供指导.灰狼优化算法具有简单高效的特点,但是在后期迭代中容易陷入局部最优.针对灰狼优化算法的缺点,对其全局优化能力进行了改进,并用改进的算法对支持向量机回归算法（SVR）的参数进行寻优,建立了MGWO-SVR预测模型.最后以中国环境监测总站中太原市的数据为研究对象,分别用MGWO-SVR模型和SVR模型对太原市的空气质量指数进行了预测拟合实验.实验结果表明,MGWO-SVR模型可以有效预测空气质量指数,并比SVR模型有更高的预测精度.The Air Quality Index（AQI）forecast can provide guidance for enterprise and social work. Grey wolf Optimizer has the characteristics of simple and efficient, but it is easy to fall into local optimum in later iteration. In this paper, the Grey wolf Optimizer is improved by using the improved algorithm to optimize the parameters of support vector machine regression algorithm （SVR）, and the MGWO-SVR prediction model is established. Finally, the air quality index of Taiyuan City was simulated by MGWO-SVR model and SVR model respectively. The experiment was carried out by using the data of Taiyuan city in China National Environmental Monitoring Centre. The experimental results show that the MGWO-SVR model can predict the air quality index effectively and has higher prediction accuracy than the SVR model.

关 键 词：空气质量预测 GWO SVR MGWO MGWO-SVR 

分 类 号：X51[环境科学与工程—环境工程]