基于PCA和支持向量机的径流预测应用研究

Application research on runoff forecast based on principal component analysis and support vector machine

影响径流量的因素很多,并且这些因素与径流量之间存在着复杂的非线性关系。将主成分分析和支持向量机相结合,首先进行特征提取,降低数据维数,获取数据的主要信息;然后利用支持向量机建立径流预测模型,取得了非常好的效果。并与支持向量机回归模型进行了比较,结果表明该方法具有更好的预测精度,值得推广。

There are many factors witch affect runoff,and there is complex nonlinear relation between these factors and runoff.Combined the principal component analysis with the support vector machine,the paper firstly extracted the feature,then reduced the data dimension and acquired the key message of the data;then the runoff forecast model was established by using the support vector machine,very good effect has been obtained.Comparison is made between this model and SVM regression model,the result indicated that this method has better prediction precision which is deserved to be popularized.