Matlab

Interface for Matlab.

Instructions for building ThunderSVM

• Please refer to Installation for building ThunderSVM.

• Then, under ./build/lib/ of the ThunderSVM root directory, you should be able to see a library of ThunderSVM (e.g., libthundersvm.so on Linux machines).

• After you have successfully done the above two steps, it is ready to start using this Matlab interface.

Methods

svm_train_matlab(parameters): The format of parameters is the same as libsvm train svm using the parameters.

svm_predict_matlab(parameters): The format of parameters is the same as libsvm run svm predict using the parameters.

Examples

n = ["-c", "10", "-g", "0.125", "../dataset/test_dataset.txt", "test_dataset.model"]
m = cellstr(n)
svm_train_matlab(m)
n = ["../dataset/test_dataset.txt", "test_dataset.model", "test_datset.out"]
m = cellstr(n)
svm_predict_matlab(m)