How to compute SFS(Sequential feature selection) with "multiple class" SVM？

Question

Hsiang-Sheng Wang 2016-4-14

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/278963-how-to-compute-sfs-sequential-feature-selection-with-multiple-class-svm

回答： Nisar Ahmed 2018-9-20

I can use sequentialfs function to compute the SFS with two-class SVM.(means the binary output [0 1])

below is my code：

%% x=features, y=binary response. c = cvpartition(y,'k',10); opts = statset('display','iter'); [fs,history] = sequentialfs(fun,x,y,'cv',c,'options',opts)

function err = SVM_class_fun(xTrain, yTrain, xTest, yTest) model = svmtrain(xTrain, yTrain, 'Kernel_Function', 'polynomial','polyorder',2, 'boxconstraint', 1); err = sum(svmclassify(model, xTest) ~= yTest); end

my issue is how to cumpute SFS when y is not binary.

I can't try it successfully,

help me, thanks！