巢湖水质安全评价的对应分析和投影寻踪熵耦合方法

在巢湖水质安全评价指标体系及其等级标准建立的基础上,把指标的客观权重分解为反映不同指标属性对评价结果的影响程度的属性权重和反映不同评价对象指标值之间的整体差异对评价结果的影响程度的样本集权重,分别用基于加速遗传算法的对应分析方法和投影寻踪方法确定,再用最小相对信息熵原理计算指标的客观组合权重,形成了水质安全评价的对应分析和投影寻踪熵耦合方法(CFA-PP)。CFA-PP的应用结果表明:影响巢湖水质安全最大的前4个指标的客观权重分别是总磷占25.5%,叶绿素a占21.6%,氨氮占18.4%,总氮占12.7%;全湖水质安全处于"较不安全"与"不安全"等级之间;控制磷、氮营养盐分的入湖负荷是提高水质安全程度的有效途径。CFA-PP是由数据驱动的综合评价方法,在水质安全评价中具有应用价值。

Entropy Coupling Method of Evaluating Chaohu Lake Water Quality Security Using Correspondence Factor Analysis and Projection Pursuit

On the basis of evaluation index system of Chaohu Lake water quality security and its grading criterion,an objective weight structure of evaluation indexes consisted of attribute weights and sample set weights is proposed,the attribute weights can be computed using accelerating genetic algorithm based correspondence factor analysis,the sample set weights can be computed using accelerating genetic algorithm based projection pursuit method,the objective combination weights of evaluation indexes can be computed using minimum relative information entropy principle,which forms a model for evaluating Chaohu Lake water quality security,named CFA-PP for short.The applied result of CFA-PP shows that the most important four indexes are total phosphorus,chlorophyll a,ammonia nitrogen and total nitrogen,respectively;the water quality security grade of the whole lake is between "relative insecurity" and "insecurity" grade;the most effective approach to improve water quality security of Chaohu lake is to reduce influent trophic salt loads of phosphorus and nitrogen.As a general evaluation method driven by data,CFA-PP can be widely applied in different water quality security evaluation.