在阐述支持向量机(SVM)和最小二乘支持向量机(LS-SVM)的原理并比较了两者的优缺点后,将LS-SVM与多层前向网络中的两种典型网络BP网络和RBF网络,分别应用于装载机载重动态测量的非线性函数回归估计中,对这三种网络在函数逼近和泛化能力两方面的性能进行比较研究。仿真结果表明,LS-SVM在精度和泛化性能两方面做到了最好的折衷,是用于非线性函数回归分析的一种很有效的方法。 After elaborating the principle of Support Vector Machine (SVM) and Least Squares Support Vector Machine (LS-SVM) and comparing the advantages and disadvantages of both, this paper compares LS-SVM with two typical network BP networks And RBF networks are respectively applied to nonlinear function regression estimation of loader load dynamic measurement. The performance of these three networks in function approximation and generalization is compared. Simulation results show that LS-SVM is the best compromise between accuracy and generalization performance, and it is a very effective method for regression analysis of nonlinear functions.