摘要
传统的数学预测模型对于小样本空气污染物数据预测的误差较大,不利于空气质量发展特征的科学分析。为优化灰色模型在空气质量预测分析中的精度,计算GM(1,1)模型、多项式回归残差修正GM(1,1)模型、PSO背景权值优化GM(1,1)模型预测值与实际值的灰色关联度,以关联度为依据使用数学方法确定各模型的权重系数,重新构建一个高精度的灰色关联组合模型,以获取各城市的PM10、PM2.5浓度预测值。河南省城市空气质量预测结果显示,该模型能够给出可视化的图表预测结果,便于研究者对区域性的空气质量发展规律进行分析探究;相比单一的灰色模型而言,该模型的预测误差小、稳定性强、可视化分析效果突出。
The traditional mathematical prediction model has a big error in predicting the data of small samples of air pollutants,which is not conducive to the scientific analysis of the development characteristics of air quality.To optimize the accuracy of grey model in air quality prediction and analysis,the grey correlation degree between the predicted value and the actual value of GM(1,1)model,polynomial regression residual correction GM(1,1)model and PSO background weight optimization GM(1,1)model is calculated.Based on the correlation degree,the weight coefficient of each model is determined by mathematical method,and a high accuracy grey correlation combination model is reconstructed to obtain PM10 and PM2.5 of each city.The prediction results of urban air quality in Henan Province show that the model can give visual chart prediction results,which is convenient for researchers to analyze and explore the regional air quality development law.Compared with the single grey model,this model has small prediction error,strong stability and outstanding visual analysis effect.
作者
郑瑶
邢昱
郭悦嵩
李明
Zheng Yao;Xing Yu;Guo Yuesong;Li Ming(Ecological Environment Monitoring Center of Henan Province,Zhengzhou 450046,China;Henan Key Laboratory of Environmental Monitoring Technology,Zhengzhou 450046,China)
出处
《能源与环保》
2022年第8期50-56,63,共8页
CHINA ENERGY AND ENVIRONMENTAL PROTECTION
基金
河南省重大科技专项(201400210700)。
关键词
灰色模型
关联度
空气质量
背景权值
grey model
correlation degree
air quality
background weight
