| 摘要: |
| 【目的】 针对地下水水质评价中影响因素的模糊性和各因素与评价等级之间的不确定性等问题，采用支持向量机模型进行地下水水质评价研究，改进地下水水质评价方式。【方法】 应用非线性支持向量机模型中的分类支持向量机，选用Gauss核函数，以羊毛湾灌区部分水井的水质资料为研究对象，进行地下水水质评价，并利用综合指数法和BP人工神经网络法对评价结果进行验证。【结果】 羊毛湾灌区水质评价结果显示，该区地下水资源已被污染，需要进行保护性开发。3种方法的评价结果较为相似，但相较于综合指数法，支持向量机计算速度较快，易于通过计算机实现；相较于标准BP人工神经网络模型，支持向量机的评价精度较高，收敛速度较快，且所需参数较少。【结论】 支持向量机能将复杂的非线性问题转化为线性问题，从而有效地避免过学习问题，并且拥有极大的泛化能力和对小样本问题的处理能力，可有效提高地下水水质评价精度，简化评价过程，为地下水水质评价提供了一条新思路。 |
| 关键词: 支持向量机 地下水 水质评价 |
| DOI： |
| 分类号: |
| 基金项目:国家科技支撑计划项目（2006BAD11B05）；国家自然科学基金项目（50879071） |
|
| Application of Support Vector Machine in the groundwater quality evaluation |
|
|
| Abstract: |
| 【Objective】Oweing to the fuzziness of the affecting factors and uncertain relationship between the factors and groundwater quality levels,the theory of Support Vector Machine(SVM) is introduced to study the groundwater evaluation,so that the groundwater quality evaluation,and the development and utilization of groundwater resources of Yangmaowan Irrigated Area could be improved.【Method】Choosing the Classified-Support Vector Machine(C-SVM) and Gauss kernel function,using the data of Yangmaowan Irrigated Area,groundwater quality is evaluated.And the result is compared with the one of Synthetical Index Method and BP Artificial Neural Network (BP ANN).【Result】From the result of groundwater quality evaluation of Yangmaowan Irrigated Area,we can see that,the groundwater should be protective development.The results from 3 different methods are similar,but comparing with Synthetical Index Method,the calculation of SVM is faster and easier to achieve on the computer;and comparing with standard BP ANN,SVM is more accurate,faster,and need fewer parameters. 【Conclusion】The method based on SVM can translate nonlinearity to linearity,so it can avoid over-fitting problem.Besides,it also has better generalization ability and handling ability with small sample data.Apparently,it can also improve the accuracy effectively and simplify the calculation.Overall,SVM is a new way to evaluate the groundwater quality. |
| Key words: SVM groundwater quality water quality evaluation |
