降水pH值的支持向量回归预测模型构建  

APPLICATION OF SUPPORT VECTOR REGRESSION TO THE PREDICTION OF pH VALUE IN PRECIPITATION

作　　者：印家健[1] 姜微[1] 戴松林 李梦龙[1] 

机构地区：[1]四川大学化学学院,成都610064 [2]东莞市环境保护监测站,东莞523011

出　　处：《环境化学》2006年第2期211-214,共4页Environmental Chemistry

摘　　要：将支持向量回归用于降水pH值预测模型的构建,结果表明,该模型具有较好的稳定性和较高的预测精度,降水的pH值主要受大气中碱性离子浓度的影响,起主导作用的是碱性离子的中和作用;其预测结果优于多元线性回归、主成分回归、偏最小二乘回归和投影寻踪回归等模型．The support vectors regression （SVR） is firstly introduced to investigate the data of the Dongguan＇s precipitation of 2003. The model was built to predict the pH value in precipitation. The result shows this method is steady and has good prediction precision and indicates pH value of Dongguan＇s precipitation mainly impacted by alkalescence ion and its neutralization action, the model has a better prediction result than that of multiple linear regression （MLR）, principal component regression （PCR）, partial least squares regression （ PLSR）, projection pursuit regression （PPR）.

关 键 词：酸雨 支持向量回归 PH值 预测 

分 类 号：X132[环境科学与工程—环境科学]