改进的主成分分析法在水库水质评价中的应用研究  

Application Research of the Improved Principal Component Analysis on Reservoir Water Quality Evaluation

作　　者：刘清园 李永[1] 蒲迅赤[1] 唐锡良 

机构地区：[1]四川大学水力学与山区河流开发保护国家重点实验室,成都610065 [2]中国三峡建设管理有限公司,成都610041

出　　处：《四川环境》2017年第6期116-122,共7页Sichuan Environment

基　　金：国家自然科学基金(51479127);国家重点研发计划"典型脆弱生态修复与保护研究"重点专项(2016YFC0502210)

摘　　要：利用SPSS软件对数据进行分析,对传统的主成分标准化法进行改进,使用改进后的主成分分析法(PCA)对α水库10个监测断面2016年7月的水质进行综合评价,结果表明用3个主成分因子可以对所有样本进行评价分析,并得到了各断面综合得分和区域综合主成分分级标准。各断面综合得分(污染程度从重到轻)为G1>G3>Z9>G11>G12>G8>G10>G5>G15>G14,分析所得各断面主要受污染因子与断面实测资料吻合,说明改进后的主成分分析方法可以在保证尽可能少丢失原始数据的基础上对水质进行合理评价,不仅避免了传统方法的主观性、随意性以及多指标数据的复杂性,还可以突出断面主要影响因子并指导控制治理措施的进行。After improving the traditional principal component standardization method, the water quality of 10 monitoring sections of ot reservoir in July 2016 were comprehensively evaluated by the improved principal component analysis （PCA） method through SPSS. The result shows that 3 principal component factors can be used to evaluate all the samples. Meanwhile, the comprehensive evaluation results of the water quality and integrated criteria for classification of principal components are obtained through calculation. The comprehensive scores of each section （ pollution degree from severe to mild） are G1 〉 G3 〉 Z9 〉 G11 〉 G12 〉 G8 〉 G10 〉 G5 〉 G15 〉 G14. The main pollution factors of each section calculated by PCA are in good agreement with the measured data. It indicates that the improved principal component analysis method can evaluate the water quality reasonably on the basis of a minirtmm loss of the original data. It not only avoids the subjectivity, randonmess of traditional methods and the complexity of multi index data, but also highlights the main influencing factors of monitoring sections and conducts the control measures.

关 键 词：SPSS软件 改进的主成分分析 水质评价 

分 类 号：X524[环境科学与工程—环境工程]