Empirical exploitation of click data for task specific ranking
A. Dong, Y. Chang, S. Ji, C. Liao, X. Li, and Z. Zheng. EMNLP '09: Proceedings of the 2009 Conference on Empirical Methods in Natural Language Processing, page 1086--1095. Morristown, NJ, USA, Association for Computational Linguistics, (2009)
Abstract
There have been increasing needs for task specific rankings in web search such as rankings for specific query segments like long queries, time-sensitive queries, navigational queries, etc; or rankings for specific domains/contents like answers, blogs, news, etc. In the spirit of "divide-and-conquer", task specific ranking may have potential advantages over generic ranking since different tasks have task-specific features, data distributions, as well as feature-grade correlations. A critical problem for the task-specific ranking is training data insufficiency, which may be solved by using the data extracted from click log. This paper empirically studies how to appropriately exploit click data to improve rank function learning in task-specific ranking. The main contributions are 1) the exploration on the utilities of two promising approaches for click pair extraction; 2) the analysis of the role played by the noise information which inevitably appears in click data extraction; 3) the appropriate strategy for combining training data and click data; 4) the comparison of click data which are consistent and inconsistent with baseline function.
Description
Empirical exploitation of click data for task specific ranking
%0 Conference Paper
%1 dong2009empirical
%A Dong, Anlei
%A Chang, Yi
%A Ji, Shihao
%A Liao, Ciya
%A Li, Xin
%A Zheng, Zhaohui
%B EMNLP '09: Proceedings of the 2009 Conference on Empirical Methods in Natural Language Processing
%C Morristown, NJ, USA
%D 2009
%I Association for Computational Linguistics
%K click data exploitation implicit-feedback learning-to-rank
%P 1086--1095
%T Empirical exploitation of click data for task specific ranking
%U http://portal.acm.org/citation.cfm?id=1699654&dl=GUIDE&coll=GUIDE&CFID=97288569&CFTOKEN=50071194
%X There have been increasing needs for task specific rankings in web search such as rankings for specific query segments like long queries, time-sensitive queries, navigational queries, etc; or rankings for specific domains/contents like answers, blogs, news, etc. In the spirit of "divide-and-conquer", task specific ranking may have potential advantages over generic ranking since different tasks have task-specific features, data distributions, as well as feature-grade correlations. A critical problem for the task-specific ranking is training data insufficiency, which may be solved by using the data extracted from click log. This paper empirically studies how to appropriately exploit click data to improve rank function learning in task-specific ranking. The main contributions are 1) the exploration on the utilities of two promising approaches for click pair extraction; 2) the analysis of the role played by the noise information which inevitably appears in click data extraction; 3) the appropriate strategy for combining training data and click data; 4) the comparison of click data which are consistent and inconsistent with baseline function.
%@ 978-1-932432-63-3
@inproceedings{dong2009empirical,
abstract = {There have been increasing needs for task specific rankings in web search such as rankings for specific query segments like long queries, time-sensitive queries, navigational queries, etc; or rankings for specific domains/contents like answers, blogs, news, etc. In the spirit of "divide-and-conquer", task specific ranking may have potential advantages over generic ranking since different tasks have task-specific features, data distributions, as well as feature-grade correlations. A critical problem for the task-specific ranking is training data insufficiency, which may be solved by using the data extracted from click log. This paper empirically studies how to appropriately exploit click data to improve rank function learning in task-specific ranking. The main contributions are 1) the exploration on the utilities of two promising approaches for click pair extraction; 2) the analysis of the role played by the noise information which inevitably appears in click data extraction; 3) the appropriate strategy for combining training data and click data; 4) the comparison of click data which are consistent and inconsistent with baseline function.},
added-at = {2010-10-01T09:09:36.000+0200},
address = {Morristown, NJ, USA},
author = {Dong, Anlei and Chang, Yi and Ji, Shihao and Liao, Ciya and Li, Xin and Zheng, Zhaohui},
biburl = {https://www.bibsonomy.org/bibtex/2cc9097a839c8e7e10c9ffcd1612040f7/beate},
booktitle = {EMNLP '09: Proceedings of the 2009 Conference on Empirical Methods in Natural Language Processing},
description = {Empirical exploitation of click data for task specific ranking},
interhash = {aef86001afda0b51d767377d4724249b},
intrahash = {cc9097a839c8e7e10c9ffcd1612040f7},
isbn = {978-1-932432-63-3},
keywords = {click data exploitation implicit-feedback learning-to-rank},
location = {Singapore},
pages = {1086--1095},
publisher = {Association for Computational Linguistics},
timestamp = {2010-10-01T09:09:37.000+0200},
title = {Empirical exploitation of click data for task specific ranking},
url = {http://portal.acm.org/citation.cfm?id=1699654&dl=GUIDE&coll=GUIDE&CFID=97288569&CFTOKEN=50071194},
year = 2009
}