一种过程支持向量机模型及其若干理论性质

A process support vector machine model and some of its theoretical properties

针对时变信号模式分类问题,建立一种过程支持向量机模型.该模型的输入为时变函数,通过核函数变换将动态模式映射到高维特征空间,经过学习训练集中函数样本类别特性,自适应提取动态模式的过程特征,直接分类辨识时变信号.证明过程支持向量机与单隐层前馈过程神经元网络的二分类能力等价;将复杂的动态模式集合非线性地映射到高维特征空间,提高动态模式的可分性;传统支持向量机是过程支持向量机的一种特例等理论性质.

Aiming at the pattern classification problem of time-varying signal, a Process Support Vector Machine (PSVM) model is presented in this paper. The inputs of PSVM can be time-varying functions. Through the kernel function transforming, dynamic pattern is mapped into high-dimensional feature space. After learning classification characteristic of the training samples, PSVM can extract process characteristics of time-varying function adaptively and classify time-varying signals directly. Some theoretical problems were proved, such as the equivalence on two-category ability of PSVM and three-layer feedforward process neural networks, complex dynamic pattern sets being nonlinearly mapped into high-dimensional feature space to improve the separability of dynamic pattern, traditional SVM is a special case of PSVM, etc.