论文部分内容阅读
在研究支持向量机(SVM)核方法和小波框架理论的基础上,提出了一种称为小波支持向量机(WaveletSupport Vector Machines,WSVM)的新的机器学习构造方法.该方法引入小波基函数构造SVM的核函数,得到了一种新的SVM模型,然后提出了此模型的结构设计和实现算法,最后给出了几种常用的小波核函数,并给出了理论证明.通过仿真实验,把该方法与小波神经网络、高斯核SVM相比较,得到了较好的实验结果,从而验证了该方法的正确性和有效性.
Based on the study of SVM kernel method and wavelet frame theory, a new machine learning method called WaveletSupport Vector Machines (WSVM) is proposed in this paper, which introduces the construction of wavelet basis function SVM kernel function, a new SVM model is obtained, and then the structural design and implementation algorithm of this model are proposed. Finally, several commonly used wavelet kernel functions are given, and theoretical proofs are given.Through simulation experiments, Compared with the wavelet neural network and the Gaussian kernel SVM, this method obtains better experimental results, which verifies the correctness and effectiveness of the proposed method.