Distance metric learning vs. Fisher discriminant analysis
A. Babak, B. Michael, and G. Ali. Proceedings of the 23rd national conference on Artificial intelligence - Volume 2, page 598--603. AAAI Press, (2008)
Abstract
There has been much recent attention to the problem of learning an appropriate distance metric, using class labels or other side information. Some proposed algorithms are iterative and computationally expensive. In this paper, we show how to solve one of these methods with a closed-form solution, rather than using semidefinite programming. We provide a new problem setup in which the algorithm performs better or as well as some standard methods, but without the computational complexity. Furthermore, we show a strong relationship between these methods and the Fisher Discriminant Analysis.
Description
Distance metric learning vs. Fisher discriminant analysis
%0 Conference Paper
%1 Alipanahi:2008:DML:1620163.1620164
%A Babak, Alipanah
%A Michael, Biggs
%A Ali, Ghodsi
%B Proceedings of the 23rd national conference on Artificial intelligence - Volume 2
%D 2008
%I AAAI Press
%K distance learning metric
%P 598--603
%T Distance metric learning vs. Fisher discriminant analysis
%U http://dl.acm.org/citation.cfm?id=1620163.1620164
%X There has been much recent attention to the problem of learning an appropriate distance metric, using class labels or other side information. Some proposed algorithms are iterative and computationally expensive. In this paper, we show how to solve one of these methods with a closed-form solution, rather than using semidefinite programming. We provide a new problem setup in which the algorithm performs better or as well as some standard methods, but without the computational complexity. Furthermore, we show a strong relationship between these methods and the Fisher Discriminant Analysis.
%@ 978-1-57735-368-3
@inproceedings{Alipanahi:2008:DML:1620163.1620164,
abstract = {There has been much recent attention to the problem of learning an appropriate distance metric, using class labels or other side information. Some proposed algorithms are iterative and computationally expensive. In this paper, we show how to solve one of these methods with a closed-form solution, rather than using semidefinite programming. We provide a new problem setup in which the algorithm performs better or as well as some standard methods, but without the computational complexity. Furthermore, we show a strong relationship between these methods and the Fisher Discriminant Analysis.},
acmid = {1620164},
added-at = {2011-10-13T10:59:59.000+0200},
author = {Babak, Alipanah and Michael, Biggs and Ali, Ghodsi},
biburl = {https://www.bibsonomy.org/bibtex/2b4e26df5e610778cb8b3f3cd9ed28139/punko},
booktitle = {Proceedings of the 23rd national conference on Artificial intelligence - Volume 2},
description = {Distance metric learning vs. Fisher discriminant analysis},
interhash = {bd1a4aa6c768c6011ce425fd7a471389},
intrahash = {b4e26df5e610778cb8b3f3cd9ed28139},
isbn = {978-1-57735-368-3},
keywords = {distance learning metric},
location = {Chicago, Illinois},
numpages = {6},
pages = {598--603},
publisher = {AAAI Press},
series = {AAAI'08},
timestamp = {2011-10-13T10:59:59.000+0200},
title = {Distance metric learning vs. Fisher discriminant analysis},
url = {http://dl.acm.org/citation.cfm?id=1620163.1620164},
year = 2008
}