Metric embedding for kernel classification rules

Bharath K. Sriperumbudur, Omer A. Lang, Gert R.G. Lanckriet

Research output: Chapter in Book/Report/Conference proceedingConference contribution

4 Citations (Scopus)

Abstract

In this paper, we consider a smoothing kernel based classification rule and propose an algorithm for optimizing the performance of the rule by learning the bandwidth of the smoothing kernel along with a data-dependent distance metric. The data-dependent distance metric is obtained by learning a function that embeds an arbitrary metric space into a Euclidean space while minimizing an upper bound on the resubstitution estimate of the error probability of the kernel classification rule. By restricting this embedding function to a reproducing kernel Hubert space, we reduce the problem to solving a semidefinite program and show the resulting kernel classification rule to be a variation of the k-nearest neighbor rule. We compare the performance of the kernel rule (using the learned data-dependent distance metric) to state-of-the-art distance metric learning algorithms (designed for k-nearest neighbor classification) on some benchmark datasets. The results show that the proposed rule has either better or as good classification accuracy as the other metric learning algorithms.

Original languageEnglish (US)
Title of host publicationProceedings of the 25th International Conference on Machine Learning
Pages1008-1015
Number of pages8
StatePublished - Nov 26 2008
Event25th International Conference on Machine Learning - Helsinki, Finland
Duration: Jul 5 2008Jul 9 2008

Publication series

NameProceedings of the 25th International Conference on Machine Learning

Other

Other25th International Conference on Machine Learning
CountryFinland
CityHelsinki
Period7/5/087/9/08

Fingerprint

Learning algorithms
Bandwidth
Error probability

All Science Journal Classification (ASJC) codes

  • Artificial Intelligence
  • Human-Computer Interaction
  • Software

Cite this

Sriperumbudur, B. K., Lang, O. A., & Lanckriet, G. R. G. (2008). Metric embedding for kernel classification rules. In Proceedings of the 25th International Conference on Machine Learning (pp. 1008-1015). (Proceedings of the 25th International Conference on Machine Learning).
Sriperumbudur, Bharath K. ; Lang, Omer A. ; Lanckriet, Gert R.G. / Metric embedding for kernel classification rules. Proceedings of the 25th International Conference on Machine Learning. 2008. pp. 1008-1015 (Proceedings of the 25th International Conference on Machine Learning).
@inproceedings{5dbc46995dc345ed9c056f32bb8fd416,
title = "Metric embedding for kernel classification rules",
abstract = "In this paper, we consider a smoothing kernel based classification rule and propose an algorithm for optimizing the performance of the rule by learning the bandwidth of the smoothing kernel along with a data-dependent distance metric. The data-dependent distance metric is obtained by learning a function that embeds an arbitrary metric space into a Euclidean space while minimizing an upper bound on the resubstitution estimate of the error probability of the kernel classification rule. By restricting this embedding function to a reproducing kernel Hubert space, we reduce the problem to solving a semidefinite program and show the resulting kernel classification rule to be a variation of the k-nearest neighbor rule. We compare the performance of the kernel rule (using the learned data-dependent distance metric) to state-of-the-art distance metric learning algorithms (designed for k-nearest neighbor classification) on some benchmark datasets. The results show that the proposed rule has either better or as good classification accuracy as the other metric learning algorithms.",
author = "Sriperumbudur, {Bharath K.} and Lang, {Omer A.} and Lanckriet, {Gert R.G.}",
year = "2008",
month = "11",
day = "26",
language = "English (US)",
isbn = "9781605582054",
series = "Proceedings of the 25th International Conference on Machine Learning",
pages = "1008--1015",
booktitle = "Proceedings of the 25th International Conference on Machine Learning",

}

Sriperumbudur, BK, Lang, OA & Lanckriet, GRG 2008, Metric embedding for kernel classification rules. in Proceedings of the 25th International Conference on Machine Learning. Proceedings of the 25th International Conference on Machine Learning, pp. 1008-1015, 25th International Conference on Machine Learning, Helsinki, Finland, 7/5/08.

Metric embedding for kernel classification rules. / Sriperumbudur, Bharath K.; Lang, Omer A.; Lanckriet, Gert R.G.

Proceedings of the 25th International Conference on Machine Learning. 2008. p. 1008-1015 (Proceedings of the 25th International Conference on Machine Learning).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - Metric embedding for kernel classification rules

AU - Sriperumbudur, Bharath K.

AU - Lang, Omer A.

AU - Lanckriet, Gert R.G.

PY - 2008/11/26

Y1 - 2008/11/26

N2 - In this paper, we consider a smoothing kernel based classification rule and propose an algorithm for optimizing the performance of the rule by learning the bandwidth of the smoothing kernel along with a data-dependent distance metric. The data-dependent distance metric is obtained by learning a function that embeds an arbitrary metric space into a Euclidean space while minimizing an upper bound on the resubstitution estimate of the error probability of the kernel classification rule. By restricting this embedding function to a reproducing kernel Hubert space, we reduce the problem to solving a semidefinite program and show the resulting kernel classification rule to be a variation of the k-nearest neighbor rule. We compare the performance of the kernel rule (using the learned data-dependent distance metric) to state-of-the-art distance metric learning algorithms (designed for k-nearest neighbor classification) on some benchmark datasets. The results show that the proposed rule has either better or as good classification accuracy as the other metric learning algorithms.

AB - In this paper, we consider a smoothing kernel based classification rule and propose an algorithm for optimizing the performance of the rule by learning the bandwidth of the smoothing kernel along with a data-dependent distance metric. The data-dependent distance metric is obtained by learning a function that embeds an arbitrary metric space into a Euclidean space while minimizing an upper bound on the resubstitution estimate of the error probability of the kernel classification rule. By restricting this embedding function to a reproducing kernel Hubert space, we reduce the problem to solving a semidefinite program and show the resulting kernel classification rule to be a variation of the k-nearest neighbor rule. We compare the performance of the kernel rule (using the learned data-dependent distance metric) to state-of-the-art distance metric learning algorithms (designed for k-nearest neighbor classification) on some benchmark datasets. The results show that the proposed rule has either better or as good classification accuracy as the other metric learning algorithms.

UR - http://www.scopus.com/inward/record.url?scp=56449095464&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=56449095464&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:56449095464

SN - 9781605582054

T3 - Proceedings of the 25th International Conference on Machine Learning

SP - 1008

EP - 1015

BT - Proceedings of the 25th International Conference on Machine Learning

ER -

Sriperumbudur BK, Lang OA, Lanckriet GRG. Metric embedding for kernel classification rules. In Proceedings of the 25th International Conference on Machine Learning. 2008. p. 1008-1015. (Proceedings of the 25th International Conference on Machine Learning).