.png){kind=icon}
.png){kind=icon}
Return
Large margin nearest local mean classifier
.png){kind=icon}
PRE
AI
DOI：10.1016/j.sigpro.2009.06.015.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Distance metric learning and classifier design are two highly challenging tasks in the machine learning community. In this paper we propose a new large margin nearest local mean (LMNLM) scheme to consider them jointly, which aims at improving the separability between local parts of different classes. We adopt 'local mean vector' as the basic classification model, and then through linear transformation, large margins between heterogeneous local parts are introduced. Moreover, by eigenvalue decomposition, we may also reduce data's dimensions. LMNLM can be formulated as a semidefinite programming (SDP) problem, so it is assured to converge globally. Experimental results show that LMNLM is a promising algorithm due to its leading to high classification accuracies and low dimensions. (C) 2009 Elsevier B.V. All rights reserved.
Keywords:
Large margin
Local mean vector
Linear transformation
Eigenvalue decomposition
Dimension
Semidefinite programming
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
IF:
3.6
Papers:
9.9K
Citations:
1.7W