Whereas search engines assist users in locating initial information sources, often an overwhelmingly large number of ULRs is returned, and the task of browsing websites rests heavily on users. The contribution of this work is developing an information filtering agent (IFA) that assists users in identifying out-of-context web pages and rating the relevance of web pages. An IFA determines the relevance of web pages by adopting three heuristics: (i) detecting evidence phrases (EP) constructed from WORDNET's ontology, (ii) counting the frequencies of EP and (iii) considering the nearness among keywords. Favorable experimental results show that the IFA's ratings of web pages are generally close to human ratings in many instances. The strength and weaknesses of the IFA are also discussed.
Beschreibung
Toward an ontology-enhanced information filtering agent
%0 Journal Article
%1 sim2004ontology
%A Sim, Kwang Mong
%C New York, NY, USA
%D 2004
%I ACM
%J SIGMOD Rec.
%K agent collaborativefiltering recommendation
%N 1
%P 95--100
%R http://doi.acm.org/10.1145/974121.974138
%T Toward an ontology-enhanced information filtering agent
%U http://portal.acm.org/citation.cfm?id=974138
%V 33
%X Whereas search engines assist users in locating initial information sources, often an overwhelmingly large number of ULRs is returned, and the task of browsing websites rests heavily on users. The contribution of this work is developing an information filtering agent (IFA) that assists users in identifying out-of-context web pages and rating the relevance of web pages. An IFA determines the relevance of web pages by adopting three heuristics: (i) detecting evidence phrases (EP) constructed from WORDNET's ontology, (ii) counting the frequencies of EP and (iii) considering the nearness among keywords. Favorable experimental results show that the IFA's ratings of web pages are generally close to human ratings in many instances. The strength and weaknesses of the IFA are also discussed.
@article{sim2004ontology,
abstract = {Whereas search engines assist users in locating initial information sources, often an overwhelmingly large number of ULRs is returned, and the task of browsing websites rests heavily on users. The contribution of this work is developing an information filtering agent (IFA) that assists users in identifying out-of-context web pages and rating the relevance of web pages. An IFA determines the relevance of web pages by adopting three heuristics: (i) detecting evidence phrases (EP) constructed from WORDNET's ontology, (ii) counting the frequencies of EP and (iii) considering the nearness among keywords. Favorable experimental results show that the IFA's ratings of web pages are generally close to human ratings in many instances. The strength and weaknesses of the IFA are also discussed.},
added-at = {2009-10-19T22:25:50.000+0200},
address = {New York, NY, USA},
author = {Sim, Kwang Mong},
biburl = {https://www.bibsonomy.org/bibtex/2d6e38f0a030ee10cb1644ff2b24929a0/wnpxrz},
description = {Toward an ontology-enhanced information filtering agent},
doi = {http://doi.acm.org/10.1145/974121.974138},
interhash = {e6402bae2f48c1f05eb0109cdc46b64e},
intrahash = {d6e38f0a030ee10cb1644ff2b24929a0},
issn = {0163-5808},
journal = {SIGMOD Rec.},
keywords = {agent collaborativefiltering recommendation},
number = 1,
pages = {95--100},
publisher = {ACM},
timestamp = {2009-10-19T22:25:50.000+0200},
title = {Toward an ontology-enhanced information filtering agent},
url = {http://portal.acm.org/citation.cfm?id=974138},
volume = 33,
year = 2004
}