函数学习的小波再生核支持向量回归模型(英文)

Support Vector Regression Model with Wavelet Reproducing Kernel for Function Learning

受小波理论与再生核Hilbert空间理论的启发,提出了一种新的小波再生核。该小波再生核由不同分辨率的小波基函数生成,并且是一种容许的支持向量核。应用该小波再生核,构造了用于函数学习的最小二乘支持向量回归模型。这种回归模型融合了支持向量机与小波的优点。仿真例子说明了该方法的可行性与有效性。

Based on the theory of wavelet analysis and reproducing kernel Hibert space(RKHS),a new wavelet reproducing kernel is proposed in this paper.The generated bywavelet basis function has different resolution,and the wavelet reproducing kernel is an admissive support vector kernel which can be used to construct least square support vector regression model for function learning.The regression model incorporates the advantage of the support vector regression and the wavelet analysis.Simulation examples are given to illustrate the feasibility and effectiveness of the method.