chapter  18
Distance Metric Learning for Data Classification
ByFei Wang
Pages 14

In this survey, we will give an overview of the existing supervised distance metric learning approaches and point out their strengths and limitations, as well as present challenges and future research directions.We focus on supervised algorithms because they are under the same setting as data classification. We will categorize those algorithms from the aspect of linear/nonlinear, local/global, transductive/inductive, and also the computational technology involved.