摘要： 支持向量机中核函数及其参数的选择具有重要意义。提出一种基于高斯核函数的支持向量机参数对快速求取方法，根据支持向量之间的几何判据，结合线性搜索法完成参数寻优，具有简单、计算量小、易于实现的优点。实验结果表明，该方法较好地解决了高斯核函数参数在实际使用中不易确定的问题，且运算速度高于原有方法。

关键词: 支持向量机, 参数优化, 高斯核函数

Abstract: It is important to determine the kernel function and its parameters for Support Vector Machine(SVM). This paper proposes a fast method based on Gauss kernel function to get SVM parameter pair. According to the geometry between support vectors, parameter optimization is accomplished by using linear search method. The method is simple and easy to program. Experimental results show the method can solve the problem that Gauss kernel function is difficult to be determined in practical application and needs less time than existing methods.

Key words: Support Vector Machine(SVM), parameter optimization, Gauss kernel function

中图分类号: 

• TP18