SVM classifier performance and flipping of arguments

2 次查看（过去 30 天）

Tahir 2015-5-9

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/215883-svm-classifier-performance-and-flipping-of-arguments

Hi, I want to calculate the performance of SVM...i follow a very useful post of this forum.. SVM classifer

Here in the last there is an argument f..i have a question about that..what this argument means?? is it necessary to flip it? What is the benefit of this flipping. here is the code...Kindly tell me about the fliping of f .

shift = svm.ScaleData.shift;
scale = svm.ScaleData.scaleFactor;
Xnew = bsxfun(@plus,Xnew,shift);
Xnew = bsxfun(@times,Xnew,scale);
sv = svm.SupportVectors;
alphaHat = svm.Alpha;
bias = svm.Bias;
kfun = svm.KernelFunction;
kfunargs = svm.KernelFunctionArgs;
f = kfun(sv,Xnew,kfunargs{:})'*alphaHat(:) + bias;
f = -f; % flip the sign to get the score for the +1 class