Wang YT, Kitsuregawa M "Link based clustering of Web search results" ADVANCES IN WEB-AGE INFORMATION MANAGEMENT, PROCEEDINGS LECTURE NOTES IN COMPUTER SCIENCE 2118: 225-236 2001

[Eugene Garfield]

{ytwang,kitsure}@tkl.iis.u-tokyo.ac.jp

Full text available at :
http://www.tkl.iis.u-tokyo.ac.jp/Kilab/Research/Paper/2001/wang/PaperforWAIM
.pdf (BE SURE TO USE THE ENTIRE URL including .pdf which is on this line)

TITLE Link based clustering of Web search results
AUTHOR Wang YT, Kitsuregawa M
JOURNAL ADVANCES IN WEB-AGE INFORMATION MANAGEMENT, PROCEEDINGS LECTURE
        NOTES IN COMPUTER SCIENCE  2118: 225-236 2001

Document type: Article     Language: English
Cited References: 21       Times Cited: 0


Abstract:
With information proliferation on the Web, how to obtain high-quality
information from the Web has been one of hot research topics in many fields
like Database, IR as well as AI. Web search engine is the most commonly used
tool for information retrieval; however, its current status is far from
satisfaction. In this paper, we propose a new approach to cluster search
results returned from Web search engine using link analysis. Unlike document
clustering algorithms in IR that based on common words/phrases shared
between documents, our approach is base on common links shared by pages
using co-citation and coupling analysis. We also extend standard clustering
algorithm K-means to make it more natural to handle noises and apply it to
web search results. By filtering some irrelevant pages, our approach
clusters high quality pages into groups to facilitate users' accessing and
browsing. Preliminary experiments and evaluations are conducted to
investigate its
effectiveness. The experiment results show that clustering on web search
results via link analysis is promising.

Author Keywords:
link analysis, co-citation, coupling, hub, authority

Addresses:
Wang YT, Univ Tokyo, Inst Ind Sci, 7-22-1 Roppongi, Tokyo, Japan
Univ Tokyo, Inst Ind Sci, Tokyo, Japan

Publisher:
SPRINGER-VERLAG BERLIN, BERLIN

IDS Number:
BT99G

ISSN:
0302-9743