Magnitude-preserving ranking algorithms

Corinna Cortes, Mehryar Mohri, Ashish Rastogi

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

Abstract

This paper studies the learning problem of ranking when one wishes not just to accurately predict pairwise ordering but also preserve the magnitude of the preferences or the difference between ratings, a problem motivated by its key importance in the design of search engines, movie recommendation, and other similar ranking systems. We describe and analyze several algorithms for this problem and give stability bounds for their generalization error, extending previously known stability results to non-bipartite ranking and magnitude of preference-preserving algorithms. We also report the results of experiments comparing these algorithms on several datasets and compare these results with those obtained using an algorithm minimizing the pairwise misranking error and standard regression.

Original languageEnglish (US)
Title of host publicationProceedings, Twenty-Fourth International Conference on Machine Learning, ICML 2007
Pages169-176
Number of pages8
Volume227
DOIs
StatePublished - 2007
Event24th International Conference on Machine Learning, ICML 2007 - Corvalis, OR, United States
Duration: Jun 20 2007Jun 24 2007

Other

Other24th International Conference on Machine Learning, ICML 2007
CountryUnited States
CityCorvalis, OR
Period6/20/076/24/07

Fingerprint

Search engines
Experiments

ASJC Scopus subject areas

  • Human-Computer Interaction

Cite this

Cortes, C., Mohri, M., & Rastogi, A. (2007). Magnitude-preserving ranking algorithms. In Proceedings, Twenty-Fourth International Conference on Machine Learning, ICML 2007 (Vol. 227, pp. 169-176) https://doi.org/10.1145/1273496.1273518

Magnitude-preserving ranking algorithms. / Cortes, Corinna; Mohri, Mehryar; Rastogi, Ashish.

Proceedings, Twenty-Fourth International Conference on Machine Learning, ICML 2007. Vol. 227 2007. p. 169-176.

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

Cortes, C, Mohri, M & Rastogi, A 2007, Magnitude-preserving ranking algorithms. in Proceedings, Twenty-Fourth International Conference on Machine Learning, ICML 2007. vol. 227, pp. 169-176, 24th International Conference on Machine Learning, ICML 2007, Corvalis, OR, United States, 6/20/07. https://doi.org/10.1145/1273496.1273518
Cortes C, Mohri M, Rastogi A. Magnitude-preserving ranking algorithms. In Proceedings, Twenty-Fourth International Conference on Machine Learning, ICML 2007. Vol. 227. 2007. p. 169-176 https://doi.org/10.1145/1273496.1273518
Cortes, Corinna ; Mohri, Mehryar ; Rastogi, Ashish. / Magnitude-preserving ranking algorithms. Proceedings, Twenty-Fourth International Conference on Machine Learning, ICML 2007. Vol. 227 2007. pp. 169-176
@inproceedings{5327bfd0182d4941a1e4056c0830979c,
title = "Magnitude-preserving ranking algorithms",
abstract = "This paper studies the learning problem of ranking when one wishes not just to accurately predict pairwise ordering but also preserve the magnitude of the preferences or the difference between ratings, a problem motivated by its key importance in the design of search engines, movie recommendation, and other similar ranking systems. We describe and analyze several algorithms for this problem and give stability bounds for their generalization error, extending previously known stability results to non-bipartite ranking and magnitude of preference-preserving algorithms. We also report the results of experiments comparing these algorithms on several datasets and compare these results with those obtained using an algorithm minimizing the pairwise misranking error and standard regression.",
author = "Corinna Cortes and Mehryar Mohri and Ashish Rastogi",
year = "2007",
doi = "10.1145/1273496.1273518",
language = "English (US)",
volume = "227",
pages = "169--176",
booktitle = "Proceedings, Twenty-Fourth International Conference on Machine Learning, ICML 2007",

}

TY - GEN

T1 - Magnitude-preserving ranking algorithms

AU - Cortes, Corinna

AU - Mohri, Mehryar

AU - Rastogi, Ashish

PY - 2007

Y1 - 2007

N2 - This paper studies the learning problem of ranking when one wishes not just to accurately predict pairwise ordering but also preserve the magnitude of the preferences or the difference between ratings, a problem motivated by its key importance in the design of search engines, movie recommendation, and other similar ranking systems. We describe and analyze several algorithms for this problem and give stability bounds for their generalization error, extending previously known stability results to non-bipartite ranking and magnitude of preference-preserving algorithms. We also report the results of experiments comparing these algorithms on several datasets and compare these results with those obtained using an algorithm minimizing the pairwise misranking error and standard regression.

AB - This paper studies the learning problem of ranking when one wishes not just to accurately predict pairwise ordering but also preserve the magnitude of the preferences or the difference between ratings, a problem motivated by its key importance in the design of search engines, movie recommendation, and other similar ranking systems. We describe and analyze several algorithms for this problem and give stability bounds for their generalization error, extending previously known stability results to non-bipartite ranking and magnitude of preference-preserving algorithms. We also report the results of experiments comparing these algorithms on several datasets and compare these results with those obtained using an algorithm minimizing the pairwise misranking error and standard regression.

UR - http://www.scopus.com/inward/record.url?scp=34547984570&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=34547984570&partnerID=8YFLogxK

U2 - 10.1145/1273496.1273518

DO - 10.1145/1273496.1273518

M3 - Conference contribution

VL - 227

SP - 169

EP - 176

BT - Proceedings, Twenty-Fourth International Conference on Machine Learning, ICML 2007

ER -