高维支持向量机的一些新发展
Some new developments of support vector machine in high dimension
-
摘要: 对高维支持向量机（SVM）的一些新发展如非凸惩罚SVM，
L_1 范数 SVM的误差界以及SVM在充分性降维中的应用进行了介绍；通过数值模拟和实例分析，展示了这些新方法在有限样本时的表现；讨论了一些可能的方向和问题．Abstract: Some selective new developments of support vector machine(SVM), including non-convex penalized SVM, the error bound of L1 norm SVM, and the application of SVM in sufficient dimension reduction are reviewed．The performance of these new methods in high-dimensional SVM is demonstrated by numerical simulation and real data analysis．Several possible new directions and issues are discussed．
下载: