Application of SVM algorithms for analysis of seawater quality

Support vector machine (SVM) algorithms were introduced to analyze the quality of seawater, and two models were constructed to analyze different seawater qualities, one is the SVM model for recognizing two kinds of seawater and the other is hierarchical support vector machines (H-SVMs) model for recognizing multi-seawater. The decision function of the first model for recognizing two kinds of seawater was applied to assess the unknown seawater samples. The trial results were consistent with the expected. It could be concluded that the parameter w in the decision functions is able to describe the weights of evaluation indices of seawater quality, which is much easier to determine the weights than fuzzy synthesis assessment (FSA). All show that SVM are based on a strict mathematical theory with a simple structure and a good generalization performance, which are worth being studied to assess seawater quality.