How to use svmpredict (libsvm)

Hey,

I'm trying to use libsvm in Matlab, but I am a bit confused regarding the svmpredict() function. The first input to it, is called testing_label_vector. As I understand it, that would be the labels of the testing data, but that doesn't really make sense, since that is what you are trying to predict. Can someone help me clarify what this input should be?

Link to the LibSVM website

[Vahid Moradi]

Hi,I need using method libsvm and explain about libsvm'parameters

Answer 1

[Jakob Sørensen]

For further reference I figured it out myself. The testing_label_vector can be added to calculate the accuracy of the prediction. If you don't have it, a zero vector of the correct size should be fine. The reported accuracy will then of course not hold though.

P.S. I know it's lame accepting my own answer, but I don't like unresolved issues :-)

[Walter Roberson]

jack mm commented "it is correct"

[Almost Artsy]

my svmpredict give me 0x0 matrix output. can you help me solve this problem?