摘要
应用小波分析和BP神经网络相结合的方法,建立大气污染物浓度预测模型。首先,利用静态小波分解将原始的大气污染物浓度序列分解为不同频段的小波系数序列;其次,将重要的气象因子和各尺度上的小波系数序列作为BP神经网络的输入;最后,对输出的各序列预测值重构,得到最终的预测结果。使用该模型对重庆市主城区某国控监测站点的PM_(10)浓度预测,结果表明,与传统的BP神经网络模型相比,该预测模型的推广能力强、预测精密度高,具有良好的应用前景。
A forecasting model of air pollutant concentration was established with a method of wavelet analy-sis and BP neural networks combining. Firstly, series of air pollutant concentration were decomposed into differ-ent frequency bands by the static wavelet decomposition. Secondly, the reconstruction series of branch of wavelet and the important meteorological factors were input into BP neural networks. Finally, the predicted results from every decomposition series were integrated as the final prediction results of the concentration. Taking some air quality monitoring sites from Chongqing as example, we predicted the PM10 concentration by the model. The re-sults show that the model had better generalization ability, higher precision of prediction and a good application prospect compared with the traditional BP neural network.
出处
《环境监测管理与技术》
CSCD
2016年第5期24-28,共5页
The Administration and Technique of Environmental Monitoring
基金
教育部留学回国人员科研启动基金资助项目(教外司留[2013]693号)
重庆市研究生教改基金资助项目(YJG43015)
