Metric learning with relative distance constraints : a modified SVM approach

C. Luo; M. Li; H. Zhang; F. Wang; Dapeng Zhang; W. Zuo

doi:10.1007/978-3-662-46248-5_30

Metric learning with relative distance constraints : a modified SVM approach

C. Luo, M. Li, H. Zhang, F. Wang, Dapeng Zhang, W. Zuo

Research output: Chapter in book / Conference proceeding › Conference article published in proceeding or book › Academic research › peer-review

1 Citation (Scopus)

Abstract

Distance metric learning plays an important role in many machine learning tasks. In this paper, we propose a method for learning a Mahanalobis distance metric. By formulating the metric learning problem with relative distance constraints, we suggest a Relative Distance Constrained Metric Learning (RDCML) model which can be easily implemented and effectively solved by a modified support vector machine (SVM) approach. Experimental results on UCI datasets and handwritten digits datasets show that RDCML achieves better or comparable classification accuracy when compared with the state-of-the-art metric learning methods.

Original language	English
Title of host publication	Intelligent Computation in Big Data Era
Publisher	Springer Berlin Heidelberg
Pages	242-249
Number of pages	8
ISBN (Print)	9783662462478
DOIs	https://doi.org/10.1007/978-3-662-46248-5_30
Publication status	Published - 2015

Keywords

Kernel method
Lagrange duality
Mahalanobis distance
Metric learning
Support vector machine

More information

10.1007/978-3-662-46248-5_30

Cite this

@inproceedings{b57bcedfb81e4f899e6e5cd515a5adff,

title = "Metric learning with relative distance constraints : a modified SVM approach",

abstract = "Distance metric learning plays an important role in many machine learning tasks. In this paper, we propose a method for learning a Mahanalobis distance metric. By formulating the metric learning problem with relative distance constraints, we suggest a Relative Distance Constrained Metric Learning (RDCML) model which can be easily implemented and effectively solved by a modified support vector machine (SVM) approach. Experimental results on UCI datasets and handwritten digits datasets show that RDCML achieves better or comparable classification accuracy when compared with the state-of-the-art metric learning methods.",

keywords = "Kernel method, Lagrange duality, Mahalanobis distance, Metric learning, Support vector machine",

author = "C. Luo and M. Li and H. Zhang and F. Wang and Dapeng Zhang and W. Zuo",

year = "2015",

doi = "10.1007/978-3-662-46248-5_30",

language = "English",

isbn = "9783662462478",

pages = "242--249",

booktitle = "Intelligent Computation in Big Data Era",

publisher = "Springer Berlin Heidelberg",

}

TY - GEN

T1 - Metric learning with relative distance constraints : a modified SVM approach

AU - Luo, C.

AU - Li, M.

AU - Zhang, H.

AU - Wang, F.

AU - Zhang, Dapeng

AU - Zuo, W.

PY - 2015

Y1 - 2015

N2 - Distance metric learning plays an important role in many machine learning tasks. In this paper, we propose a method for learning a Mahanalobis distance metric. By formulating the metric learning problem with relative distance constraints, we suggest a Relative Distance Constrained Metric Learning (RDCML) model which can be easily implemented and effectively solved by a modified support vector machine (SVM) approach. Experimental results on UCI datasets and handwritten digits datasets show that RDCML achieves better or comparable classification accuracy when compared with the state-of-the-art metric learning methods.

AB - Distance metric learning plays an important role in many machine learning tasks. In this paper, we propose a method for learning a Mahanalobis distance metric. By formulating the metric learning problem with relative distance constraints, we suggest a Relative Distance Constrained Metric Learning (RDCML) model which can be easily implemented and effectively solved by a modified support vector machine (SVM) approach. Experimental results on UCI datasets and handwritten digits datasets show that RDCML achieves better or comparable classification accuracy when compared with the state-of-the-art metric learning methods.

KW - Kernel method

KW - Lagrange duality

KW - Mahalanobis distance

KW - Metric learning

KW - Support vector machine

U2 - 10.1007/978-3-662-46248-5_30

DO - 10.1007/978-3-662-46248-5_30

M3 - Conference article published in proceeding or book

SN - 9783662462478

SP - 242

EP - 249

BT - Intelligent Computation in Big Data Era

PB - Springer Berlin Heidelberg

ER -

Metric learning with relative distance constraints : a modified SVM approach

Abstract

Keywords

More information

Fingerprint

Cite this