中文说明：
支持向量机和BP神经网络虽然都可以用来做非线性回归，但它们所基于的理论基础不同，回归的机理也不相同。支持向量机基于结构风险最小化理论，普遍认为其泛化能力要比神经网络的强。为了验证这种观点，本文编写了支持向量机非线性回归的通用Matlab程序和基于神经网络工具箱的BP神经网络仿真模块，仿真结果证实，支持向量机做非线性回归不仅泛化能力强于BP网络，而且能避免神经网络的固有缺陷——训练结果不稳定。
English Description：
Although support vector machine and BP neural network can be used to do nonlinear regression, they are based on different theoretical basis, and the regression mechanism is also different. Support vector machine is based on the theory of structural risk minimization, and its generalization ability is generally considered to be better than that of neural network. In order to verify this point of view, this paper compiles the general matlab program of support vector machine nonlinear regression and BP neural network simulation module based on neural network toolbox. The simulation results show that support vector machine nonlinear regression not only has stronger generalization ability than BP network, but also can avoid the inherent defect of neural network - unstable training results.