摘要
由于湖泊富营养化程度影响因素多,评价因素与富营养化等级之间关系复杂而且具有非线性特征。支持向量机是由Vapnik等人提出的建立在统计学习理论基础上的一种新的机器学习方法,由于其使用结构风险最小化原则代替经验风险最小化原则,解决了一些神经网络遗留的问题,又由于其应用了核函数思想,它可以较好地解决非线性问题,利用支持向量机多类分类算法,构建巢湖富营养化程度评价模型,取得较好的结果。
The level of a lake's eutrophication is affeccted by many factors,the relationship between assessment indexs and the level of eutrophication is often compicated with non-linear features.The support vector machine,which is a kind of new machine learning methods based on statistical learning theory,and put forward by Vapnik and his fellows,it brings the concept such as structural risk minimization principle instead of experimental ones,therefore,it solved some problems whick neural network can't,plus,it also uses the conception of kernel fuctions,whick can work out the non-linear problems efficacious,the paper construct the model of the level of eutrophication of Chaohu lake by using of the toolbox the support vector machine,and the results match the reality well.
出处
《环境科学与管理》
CAS
2011年第5期181-184,共4页
Environmental Science and Management
关键词
支持向量机
核函数
富营养化
supports vector machine
kernel fuction
eutrophication