Large Margin Rank Boundaries for Ordinal Regression
Advances in Large Margin Classifiers, page 115-132. Cambridge, MA, MIT Press, (2000)Abstract
Based on ideas from SV classification an algorithm is designed to
obtain Large Margin Rank Boundaries for Ordinal Regression. In other
words, a SV algorithm for learning preference relations. In addition
to that, the paper contains a detailed derivation of the corresponding
cost functions, risk functionals, and proves uniform convergence
bounds for the setting. Experimental evidence shows the good performance
of their distribution independent approach.
Description
KDubiq Blueprint
