水质评价中的冗余理论与去余算法

Redundancy Theory and Algorithm in the Water Quality Assessment

水质评价在算法层面上解决的问题是,通过实现指标隶属度到目标隶属度转换确定目标水体的污染等级;存在的实质性问题是,如何界定隶属度转换不是简单的线性转换以及如何解决隶属度转换的非线性计算.为此,通过计算由隶属度向量表征的分类信息的信息熵、确定指标对水体分类所做贡献的大小;借助区分权概念揭示指标隶属度中包含对目标水体分类的冗余值.以"冗余"为切入点,建立以冗余定理、非线性转换等定理推论为基本内容的冗余理论.用冗余理论界定隶属度转换的非线性,建立基于区分权滤波的去余算法实现隶属度转换.所建模型在指标隶属度只取"0或1"两个数值时,将退化为通常的"加权平均"线性模型.

The paper points out that at the algorithm level water quality evaluation is to solve the problem of determining the class of pollution of the target of water by conversing from index membership degree to the objective membership degree. While the substantive issue is how to defy that the conversion of membership degree is not simple linear conversion and how to solve the nonlinear computation of conversion of membership degree. According to this problem, the paper suggests confirming the contribution the index made to water classification by calculating the information entropy of classification information characterized by the membership vector and revealing that the index membership degree contains redundancy value of the classification of the target of water by using the conception of distinguish right.. Regard redundancy as cutting point, the paper aims at constructing redundancy theory with redundancy theory, nonlinear conversion and correlative theorem deduction as its main content. It defies the nonlinear of conversion of membership degree by redundancy theory and constructs redundancy removed algorithm basing on distinguish right to achieve conversion of membership degree. When the index membership degree takes only ＂0 or 1＇ two values, the model will degenerate into the usual linear model of the weighted average model.