Distance metric learning by knowledge embedding

摘要

This paper presents an algorithm which learns a distance metric from a data set by knowledge embedding and uses the new distance metric to solve nonlinear pattern recognition problems such a clustering.