Coastal water quality evaluation method based on two-classification support vector machines and particle swarm algorithm

Abstract

The invention discloses a coastal water quality evaluation method based on two-classification support vector machines and a particle swarm algorithm. The coastal water quality evaluation method comprises the following steps of: S1, selecting a coastal water quality evaluation factor as feature information, constructing a sample data set, normalizing the feature information, preprocessing the feature information, and combining all feature information to form a feature vector; S2, determining a coastal water quality evaluation level, performing binary encoding on the coastal water quality evaluation level, and constructing a support vector machine network; S3, performing parameter optimization by the particle swarm algorithm, and obtaining an optimal vector parameter; S4, training each two-classification support vector machine in the support vector machine network according to the optimal vector parameter; and S5, inputting the trained support vector machine network into a sample set to be classified, obtaining a forecast result, and evaluating the level of coastal water quality. By the coastal water quality evaluation method, levels of various types of coastal water quality can be evaluated according to pollutant features of a selected coastal domain.

Description

technical field [0001] The invention belongs to the technical field of pattern recognition and quality evaluation, and in particular relates to an evaluation method of offshore water quality based on a second-class support vector machine and a particle swarm algorithm. Background technique [0002] The ocean is an important water resource for the survival of the earth, but it is currently facing more and more serious water pollution problems. As an important content of environmental quality assessment, water environment quality assessment is one of the important means of environmental management. For the methods of comprehensive evaluation of seawater quality, early scholars at home and abroad mostly used methods such as comprehensive index method, fuzzy comprehensive evaluation method, fuzzy clustering method, etc., but these methods must presuppose some parameters of the model or subjective regulations: such as fuzzy comprehensive evaluation The weight of water quality pa...

AI Technical Summary

Problems solved by technology

For the methods of comprehensive evaluation of seawater quality, early scholars at home and abroad mostly used methods such as comprehensive index method, fuzzy comprehensive evaluation method, fuzzy clustering method, etc., but these methods must presuppose some parameters of the model or subjective regulations: such as fuzzy comprehensive evaluation The weight of the water quality parameters should be given in the method; the membership function should be given in the fuzzy cluster analysis, etc., and the evaluation results are highly subjective.

In recent years, both at home and abroad are trying to explore more reasonable evaluation methods. Among them, the artificial neural network evaluation method is widely used, but there are two problems in using it to establish a water quality evaluation model: insufficient training samples and insufficient calculation accuracy

Due to the "overfitting" phenomenon due to the large neural network structure, the model established in this way cannot guarantee its good generalization ability and predictive ability, so a more accurate water quality evaluation model is urgently needed

Images