SVM和Elman神经网络在乌鲁木齐污染物预报中的应用研究

选用2012年11月1日-2013年1月31日的逐6 h的空气污染物（SO2、NO2、PM10）和实况气象要素（温度、湿度、能见度、风速和气压）资料,利用支持向量机和Elman神经网络方法建立空气污染物预报模型。结果表明,支持向量机和Elman神经网络方法都可以得到较为理想的预测结果,支持向量机在泛化能力方面具有显著优势,预测结果更加准确。

Prediction Method Research of air pollution in Urumqi using support vector machine and Elman neural network

Based on the observed air pollutant （SO2, NO2 and PMm） data and the meteorologicalobservational data （temperature, humidity, visibility, wind speed and pressure） from November 1st,2011 to December 31st,2012,the result is found that there are obvious linear correlation betweenmeteorological elements and pollutant concentration. This paper presents the method of 6-hoursshort-time air pollution forecasting by Support Vector Machine（SVM） and Elman Network method,and the forecast results are verified. The predicted results by Support Vector Machine was superiorto that by Elman Network. The SVM＇s generalization ability has remarkable advantage,and theSVM method used for air pollutant concentration forecast has higher accuracy.