基于模糊sigmoid核的支持向量机回归建模

支持向量机中对核函数的要求为对称的半正定矩阵.来自于神经网络的sigm o id核函数在其参数满足一定条件时才成为半正定矩阵,但是这种核函数在SVM中却有很多成功的应用.本文将sigm o id核函数与模糊逻辑相结合并使其模糊化,从而简化了SVM的计算并便于用硬件实现.通过对混沌时间序列预测以及图像去噪滤波器两个实例的实验研究发现,使用模糊sigm o id核函数可以使SVM回归建模在损失较小精度的代价下,较大地降低平均CPU执行时间,便于硬件实现

Support vector regression based on fuzzy sigmoid kernel

In the support vector machines(SVM) framework,kernel function must meet certain requirements to be symmetric and positive semi-definite(PSD) matrix.Although sigmoid function derived from neural network can become a PSD kernel for proper combinations of its free parameters,it has been used in several practical and successful cases.The sigmoid kernel is combined with fuzzy logic methodology first,which makes the computation of SVM simple and of ease implementation in hardware.Experiments for chaotic time series prediction and image filter are also carried out to show that the average CPU time can be decreased markedly in favor of hardware implementation,in spite of small decrease in prediction precision.