太湖水质指标主成分分析

Principal component analysis of water quality indexes of Taihu Lake

用主成分分析的方法研究了太湖各水质监测站监测的水质指标,从原始数据出发提取了占总方差80%以上的两个主成分,对此作出了合理的解释:第一主成分主要是由水体中的N,P和有机物所引起的;而第二个主成分主要体现了水体的清澈程度。得到了水质监测站的水质分类图,从水质分类图中明显看出各站水质的污染情况及其污染原因,将大量抽象的数据变为形象的图表,降低分析的难度。

The water quality indexes for each water quality monitoring station of the Taihu Lake are studied by use of the principal component analysis method. Two principal components, which account over 80% of the total variance are extracted from original data, and some reasonable explanations are given as follows: the first principal component is mainly caused by nitrogen, phosphor, and organic matters; the second principal component mainly reflects the limpid degree of water bodies. By analysis of the two principal components, the classifying graphs of water quality of different monitoring stations are obtained, and from the graphs, it is easy to evaluate the water pollution and identify the pollution source. The principal component method turns large amount of nonobjective data into graphs and tables, so it simplifies the data analysis.