多项式与SVM预测模型的理论分析及应用比较

Theoretic analysis and application comparison between the polynomial model and the support vector machine prediction model

标准支持向量机(SVM)及其改进形式的最小二乘支持向量机(LS-SVM)基于结构风险最小化,成功解决了多项式模型在预测方面所面临的问题;文章首先从理论上分析了SVM模型比多项式回归模型在预测方面更具有优越性;具体实验结果表明,SVM模型预测精度高,抗干扰能力强,更适合在预测方面的应用。

The standard support vector machine（SVM） and its advanced form--the least squares support vector machine （LS-SVM） are based on the principle of structural risk minimization. The prediction model based on the LS-SVM can successfully solve the problems that the polynomial prediction moded encounters. The paper analyzes the superiority of the SVM model over the polynomial model theoretically. Experiment is also made. The experiment result shows that＇the SVM model is more accurate and more robust in noise resistance, and thus more suitable for prediction.