基于指标变换值的主成分投影法在水质评价中的应用

Application of Principal Component Projection in Water Quality Evaluation Based on Index Transform Values

针对水环境系统中水质指标较多,容易出现信息重叠,评价单元在不同时间或空间整体比较和排序,各项指标赋权需要构造众多评价函数等问题的基础上,采用指标变换值的主成分投影法对大连市重点饮用水源地水质进行评价,并通过单因子评价法和灰色关联分析的评价结果作对比验证。结果表明,主成分投影法与单因子评价法、灰色关联分析评价级别不同的测点在英那河、松树、大西山上,主成分投影法在这3个断面的评价级别为Ⅳ、Ⅱ、Ⅲ,灰色关联分析的为Ⅴ、Ⅲ、Ⅲ,单因子评价法的为Ⅴ、Ⅲ、Ⅳ级,主成分投影法的评价结果与其他两种评价方法相比还具有一定差异性。为此采用蒙特卡罗方法在考虑随机观测误差下对评价结果进行修正。修正后碧流河、英那河、卧龙、松树、洼子店、北大河、大西山、鸽子塘、朱隈子这9个监测点的水质级别分别为Ⅴ、Ⅴ、Ⅴ、Ⅲ、Ⅴ、Ⅴ、Ⅳ、Ⅴ、Ⅴ。

Aimed at confusing water quality indices which result in information overlap,overall comparison and ordering of evaluation units in different times or spaces,and various index weights which need to construct corresponding evaluation functions,the method for the principal component projection of index transform values was adopted to evaluate the quality of drinking water sources of Dalian,and the evaluation results of the single factor evaluation method and the gray correlation analysis were compared and verified.The results show that the evaluation results of the principal component projection method were distinct from those of the single factor evaluation method and the grey correlation analysis in evaluating water quality at the points of different levels.And the results corrected in view of random observation errors showed that the water quality levels of the nine monitoring points of Biliuhe,Yingnahe,Wolong,Songshu,Wazidian,Beidahe,Daxishan,Gezitang,Zhuweizi were Ⅴ,Ⅴ,Ⅴ,Ⅲ,Ⅴ,Ⅴ,Ⅳ,Ⅴ and Ⅴ,which were close to facts.