【摘要】 核选择是支持向量机研究中的核心问题之一,不同的核函数将产生不同的分类效果。研究了核参数和误差惩罚参数对支持向量机推广能力的影响,然后根据局部核函数与全局核函数的各自优点,提出了一种新的自适应组合核函数,并将该核函数应用于支持向量机中。最后,利用该自适应核进行不同领域数据的实验,实验结果表明由该核函数建立的支持向量机具有更好的预测能力。更多还原

【Abstract】 The kernel function selection is one of the key problems for support vector machines (SVM) and different kernel function can produce different SVM. In this paper, the influences of the kernel parameters and error penalty parameter on support vector machines’ generalization ability are researched. Then a novel adaptive compound kernel function is proposed by taking advantage of local kernel function and global kernel function. At last, the kernel function is applied in SVM and compared with other kernels in different fields experiments. The experimental results show that the novel kernel can achieve better performance than other kernels.更多还原