/Support-Vector-machine-with-Truncated-Pinball-loss-for-Multi-class-probability-estimation

the code for the paper “ support vector machine with Truncated Pinball loss for Multi-class probability estimation”

Primary LanguageMATLAB

The working environment is MATLAB, to run this code, users also need to install Tomlab (https://tomopt.com/tomlab/).

The data saved in the training data should be a 2-D matrix with the first column as the classes and the next should be the variables, and the last few columns are the probability or pseudoprobability .

The data format in the test data should be a 2-D matrix which is the same as training data.

To run the code, simply run t_pin_svm.m file.