Recursive Feature Extraction for Ordinal Regression

摘要

Most existing algorithms for ordinal regression usually seek an orientation for which the projected samples are well separated, and seriate intervals on that orientation to represent the ranks. However, these algorithms only make use of one dimension in t