摘要
核函数技术是机器学习领域应用广泛且非常有效的方法,采用核函数技术可以有效地解决在高维空间运算时遇到的维数灾难问题,不仅大大减少了在输入空间中的计算量还能够有效改善学习机的分类性能,核函数的选择以及核函数的构造一直是机器领域非常重要的问题,然而这方面的研究成果并不多。论文首先阐述了支持向量机的理论以及核函数的基本原理,介绍了目前应用比较广泛的核函数类型,考虑到局部核函数和全局核函数的优缺点并将两者结合组成新的核函数,使用改进的网格搜索法对构造核函数进行参数和组合系数进行寻优。最后将该算法应用到ORL人脸数据库中,验证了混合核函数SVM人脸分类识别效果明显优于单一核函数分类效果,实验结果证实了该算法的有效性。
The kernel function technique is a widely used and very effective method in the field of machine learning. The kernel function technique can effectively solve the dimensionality problem encountered in the high dimensional space operation,which not only greatly reduces the computational complexity in the input space. It is very important to improve the classification performance of the learning machine,the selection of the kernel function and the construction of the kernel function. However,the research results in this field are not many. This paper first introduces the theory of support vector machine and the basic principle of kernel function. It introduces the type of kernel function which is widely used at present,taking into account the advantages and disadvantages of local kernel function and global kernel function and combining the two to form a new kernel function. The improved kernel search method is used to optimize the parameters and the combination coefficients of the construction kernel function. Finally,the algorithm is applied to the ORL face database to verify that the hybrid kernel function SVM face classification recognition effect is better than the single kernel function classification effect,the experimental results confirm the effectiveness of the algorithm.
作者
马田
吴陈
乔雯雯
王元甲
沙阳阳
MA Tian;WU Chen;QIAO Wenwen;WANG Yuanjia;SHA Yangyang(School of Computer Science,Jiangsu University of Science and Technology,Zhenjiang 212000)
出处
《计算机与数字工程》
2019年第6期1338-1341,共4页
Computer & Digital Engineering
关键词
核函数
支持向量机
网格搜索
人脸识别
kernel function
support vector machine
grid search
face recognition