Matlab toolbox: ML2S-SVM, MTLSSVM

Shuo Xu and Xin An, 2019. ML2S-SVM: Multi-Label Least-Squares Support Vector Machine Classifiers. The Electronic Library, Vol. 37, No. 6, pp. 1040-1058.

The parameters are tuned with 10-fold cross validation, and then Train and test on the specific data.

load scene
[gamma_best, mu_best, p_best, loss_best] = GridML2SSVM(train_inst, train_lbl, 10, 0, 0, 0, inf); 
[alpha, b] = ML2SVMTrain(train_inst, train_lbl, gamma_best, mu_best, p_best);
[predict_lbl]= ML2SSVMPredict(test_inst, train_inst, train_lbl, alpha, b, mu_best, p_best); 

load emotions
[gamma_best, mu_best, p_best, loss_best] = GridML2SSVM(train_inst, train_lbl, 10, 0, 0, 0, inf); 
[alpha, b] = ML2SVMTrain(train_inst, train_lbl, gamma_best, mu_best, p_best);
[predict_lbl]= ML2SSVMPredict(test_inst, train_inst, train_lbl, alpha, b, mu_best, p_best); 

load yeast
[gamma_best, mu_best, p_best, loss_best] = GridML2SSVM(train_inst, train_lbl, 10, 0, 0, 0, inf); 
[alpha, b] = ML2SVMTrain(train_inst, train_lbl, gamma_best, mu_best, p_best);
[predict_lbl]= ML2SSVMPredict(test_inst, train_inst, train_lbl, alpha, b, mu_best, p_best); 

load bibtex
[gamma_best, mu_best, p_best, loss_best] = GridML2SSVM(train_inst, train_lbl, 10, 0, 0, 0, inf); 
[alpha, b] = ML2SVMTrain(train_inst, train_lbl, gamma_best, mu_best, p_best);
[predict_lbl]= ML2SSVMPredict(test_inst, train_inst, train_lbl, alpha, b, mu_best, p_best); 

The parameters are tuned with 10-fold cross validation, and then train and test on the specific data.

load scene
[gamma_best, lambda_best, p_best, loss_best] = GridMTLSSVC(train_inst, train_lbl, 10, 0, 0, 0, inf);
[alpha, b] = ML2SVMTrain(train_inst, train_lbl, gamma_best, mu_best, p_best);
[predict_lbl]= ML2SSVMPredict(test_inst, train_inst, train_lbl, alpha, b, mu_best, p_best); 

load emotions
[gamma_best, lambda_best, p_best, loss_best] = GridMTLSSVC(train_inst, train_lbl, 10, 0, 0, 0, inf);
[alpha, b] = ML2SVMTrain(train_inst, train_lbl, gamma_best, mu_best, p_best);
[predict_lbl]= ML2SSVMPredict(test_inst, train_inst, train_lbl, alpha, b, mu_best, p_best); 

load yeast
[gamma_best, lambda_best, p_best, loss_best] = GridMTLSSVC(train_inst, train_lbl, 10, 0, 0, 0, inf);
[alpha, b] = ML2SVMTrain(train_inst, train_lbl, gamma_best, mu_best, p_best);
[predict_lbl]= ML2SSVMPredict(test_inst, train_inst, train_lbl, alpha, b, mu_best, p_best); 

load bibtex
[gamma_best, lambda_best, p_best, loss_best] = GridMTLSSVC(train_inst, train_lbl, 10, 0, 0, 0, inf);
[alpha, b] = ML2SVMTrain(train_inst, train_lbl, gamma_best, mu_best, p_best);
[predict_lbl]= ML2SSVMPredict(test_inst, train_inst, train_lbl, alpha, b, mu_best, p_best);