[Reading] Support vector learning for ordinal regression

This paper extends the SVM framework for the ordinal regression (classification of ordered classes). The algorithm is easy to understand after the problem is formulated. The key contribution of the paper is showing that the ordinal regression problem can be transformed into regular SVR by using the difference of the feature vectors as the new feature vector. Also the error bound of the algorithm is presented.

I did not know much about SVM and hope I can find some useful online materials to learn it by myself in the future....