|
Title:
|
INCREASING RELEVANCE OF INTERNET SEARCH RESULTS USING A TOPIC NETWORK |
|
Author(s):
|
Stephen C F Chan , Terry C H Lai , Edward K F Dang , Michael M K Chan , Cane W K Leung |
|
ISBN:
|
972-8924-16-X |
|
Editors:
|
Pedro Isaías, Maggie McPherson and Frank Bannister |
|
Year:
|
2006 |
|
Edition:
|
2 |
|
Keywords:
|
Topic network, belief network, relevance feedback, information retrieval, internet searching |
|
Type:
|
Short Paper |
|
First Page:
|
280 |
|
Last Page:
|
284 |
|
Language:
|
English |
|
Cover:
|
|
|
Full Contents:
|
click to dowload 
|
|
Paper Abstract:
|
Because topic directories provide both a logical decomposition and a hierarchical representation of semantic relations
among topics, a growing number of search engines are incorporating them to return more relevant search results to users
queries. If users, however, prefer to focus the search to a more specific search scope, e.g. a particular topic, most of these
search engines return web pages deemed popular only in that topic. The results returned are therefore confined to a
relatively narrow search scope, and more importantly, the quality of such results depends on users knowledge about their
search domains. In an attempt to overcome such limitations, we propose a probabilistic approach to searching not only
the query topic, but also related ones. The crux of the approach is to capitalize on cross-topic relations and to generate a
topic network by utilizing a belief network. We show how the approach can effectively operate on the Yahoo! taxonomy
and further improve the quality of search results derived from incremental relevance feedback data provided by users. |
|
|
|
|
|
|
