论文部分内容阅读
在综合序列前向选择(sequential forward selection,SFS)方法和广义序列前向选择(generalized sequential forward selection,GSFS)方法的基础上,提出了基于分类精度的特征选取(sequential forward selection based on classification accuracy,CA-SFS)方法。它依次改变GSFS方法中的r值,并以支持向量机(support vector machine,SVM)作为分类器,将得出的分类精度作为准则函数对特征进行取舍。仿真实验表明CA-SFS算法不但选择了较少的特征,而且取得了较好的分类效果。
Based on the combination of sequential forward selection (SFS) and generalized sequential forward selection (GSFS) methods, the paper presents a new method based on classification accuracy, CA-SFS) method. It in turn changes the value of r in the GSFS method, and uses the SVM (Support Vector Machine) as a classifier. The resulting classification accuracy is used as a criterion function to select the features. Simulation results show that the CA-SFS algorithm not only selects fewer features but also obtains better classification results.