There has been recent interest in collecting user or assessor preferences, rather than absolute judgments of relevance, for the evaluation or learning of ranking algorithms. Since measures like precision, recall, and DCG are defined over absolute judgments, evaluation over preferences will require new evaluation measures that explicitly model them. We describe a class of such measures and compare absolute and preference measures over a large TREC collection.
%0 Conference Paper
%1 1390451
%A Carterette, Ben
%A Bennett, Paul N.
%B SIGIR '08: Proceedings of the 31st annual international ACM SIGIR conference on Research and development in information retrieval
%C New York, NY, USA
%D 2008
%I ACM
%K clickdata ir pairwise preferences ranking search
%P 685--686
%R http://doi.acm.org/10.1145/1390334.1390451
%T Evaluation measures for preference judgments
%U http://portal.acm.org/citation.cfm?id=1390451
%X There has been recent interest in collecting user or assessor preferences, rather than absolute judgments of relevance, for the evaluation or learning of ranking algorithms. Since measures like precision, recall, and DCG are defined over absolute judgments, evaluation over preferences will require new evaluation measures that explicitly model them. We describe a class of such measures and compare absolute and preference measures over a large TREC collection.
%@ 978-1-60558-164-4
@inproceedings{1390451,
abstract = {There has been recent interest in collecting user or assessor preferences, rather than absolute judgments of relevance, for the evaluation or learning of ranking algorithms. Since measures like precision, recall, and DCG are defined over absolute judgments, evaluation over preferences will require new evaluation measures that explicitly model them. We describe a class of such measures and compare absolute and preference measures over a large TREC collection.},
added-at = {2010-03-04T13:15:48.000+0100},
address = {New York, NY, USA},
author = {Carterette, Ben and Bennett, Paul N.},
biburl = {https://www.bibsonomy.org/bibtex/2f3f307daf77da084bd2cde7939f6f65b/beate},
booktitle = {SIGIR '08: Proceedings of the 31st annual international ACM SIGIR conference on Research and development in information retrieval},
description = {Evaluation measures for preference judgments},
doi = {http://doi.acm.org/10.1145/1390334.1390451},
interhash = {ccfc79aa5f5b32109cfc87f2cd08741a},
intrahash = {f3f307daf77da084bd2cde7939f6f65b},
isbn = {978-1-60558-164-4},
keywords = {clickdata ir pairwise preferences ranking search},
location = {Singapore, Singapore},
pages = {685--686},
publisher = {ACM},
timestamp = {2010-03-04T13:15:48.000+0100},
title = {Evaluation measures for preference judgments},
url = {http://portal.acm.org/citation.cfm?id=1390451},
year = 2008
}