基于逐步判别分析法的某库区水质评价

Water quality evaluation of a reservoir area based on gradual discriminant analysis

利用逐步判别分析法对水质评价指标进行筛选后,根据国家地表水质量标准随机生成305个训练样本,建立判别函数,对长江流域某库区8个监测断面2008—2012年逐月的水质进行判别评价。结果显示,选取的7项水质指标均对水质评价结果有显著影响;训练样本误判率为0.328%,检验样本评判正确率为100%,表明判别模型具有很高的判别能力;某库区水质整体情况良好,水质等级以《地表水环境质量标准》(GB 3838—2002)规定的Ⅲ类为主,主要分布在Ⅱ类、Ⅲ类、Ⅳ类。

After using stepwise discriminant analysis to screen water quality evaluation indicators,305 training samples were randomly generated according to the national surface water quality standards,and a discriminant function was established to evaluate the water quality of 8 monitoring sections in a certain reservoir area of the Yangtze River Basin from 2008 to 2012 on a monthly basis.The results showed that all 7 selected water quality indicators had a significant impact on the water quality evaluation results;the misjudgment rate of the training samples is 0.328%,and the accuracy rate of the test sample evaluation is 100%,indicating that the discriminative model has high discriminative ability;the overall water quality of a certain reservoir area is good,with the water quality grade mainly classified as class III according to the Environmental Quality Standards for Surface Water(GB 3838—2002),mainly distributed in class II,III,and IV.