Water quality assessment method based on improved rough set-set pair analysis

Abstract

The invention discloses a water quality assessment method based on improved rough set-set pair analysis. The method comprises the following steps: firstly, a three-element connection number of the set pair analysis theory is extended to a six-element connection number which corresponds to corresponding grades, the connection degree of each sample is preliminarily calculated, the connection degree of the actually measured index value of each assessment index relative to the index grading standard is calculated, the importance degree and weight of each assessment index are calculated according to conditional information entropy of an improved rough set, the weight is combined with the connection degree of each assessment index, the connection degree of each assessment sample is obtained, finally, components are subjected to normalization processing, the average connection degree, shown in the specification, of the samples is obtained, and the grade represented by the maximum value is taken as the nutrition grade of the samples. The conditional information entropy of the improved rough set is combined with set pair analysis, so that influence of human factors on weight determination can be eliminated, worry about weight being 0 can be eliminated, and the method has rationality and effectiveness by integrating various indexes for assessment.

Description

technical field [0001] The invention belongs to the technical field of water quality evaluation, in particular to a water quality evaluation method based on an improved rough set-cloud model. Background technique [0002] Industrial wastewater, domestic sewage and other wastes enter rivers, lakes, seas and other water bodies, exceeding the self-purification capacity of the water body, leading to changes in its physical, chemical, biological and other characteristics, thereby affecting the use value of the water body and restricting the sustainable development of the social economy ; at the same time, natural processes including erosive changes, weathering of the earth's crust, etc. can also reduce the use of water bodies for drinking, industry, agriculture, recreation and other purposes. Therefore, the quality evaluation of water bodies is a basic task for the rational development, utilization and protection of water resources, and an important guarantee for the sustainable ...

AI Technical Summary

Problems solved by technology

[0003] 1. Analytic Hierarchy Process (AHP) can stratify the decision-making thinking of complex systems, organically combine qualitative and quantitative factors in the decision-making process, and simplify complex problems, but there are many qualitative components and less quantitative data. not convincing

The scheme ranking and scientific decision-making with the AHP method are obtained by pairwise comparison. For some factors, when the opinions of the experts are inconsistent, it is impossible to establish a complete judgment and proof, and the judgment matrix does not have absolute consistency.

[0004] 2. The gray system evaluation method can be used to analyze imprecise data, short samples and incomplete hydrological data, but the variable values ​​must be standardized before the clustering process, that is, to eliminate the influence of dimensions, different methods of standardization will lead to different clustering results

[0005] 3. The fuzzy mathematical evaluation method can effectively solve the fuzzy boundary problem and control the impact of monitoring errors on the evaluation results. However, when determining the weight of each index factor, it is usually by using the basic experience and knowledge of people, which is subjective

[0006] 4. The artificial neural network method has strong adaptability and objective evaluation results, but it has high requirements for training samples, complex implementation process, and limitations

BP neural network and RBF neural network are two widely used network models. BP neural network is prone to local minimization, slow convergence, and different network structure choices; when RBF neural network has insufficient data, it will Doesn't work and relies too much on sample data

Images