TOPIC
Measuring Web Page Reputation
Overview
TOronto Page
Influence Computation (TOPIC)
is a prototype built around the idea that the linkage structure between
pages can be mined to extract reputation measures for Web pages or
sites. The input to the system is the URL of a Web page (or a site) and the
output is a ranked list of topics on which the page has reputation.
To provide a reasonable response time, the system examines a fixed
number of links from a larger set of links that can affect the reputation
of a page. You may change this number to a smaller number
(to get better speed) or a larger number (to get better precision).
You can also enter an optional term or phrase, in which case the
reputation of the page is measured with better precision only on that
particular topic; the result in this case is displayed within a list of
top 10 authorities identified by
Google
for the same topic, providing a comparative ranking.
Project Members
Demo
This is a snapshot of the page when it was working.
Publications/Presentations
Examples
-
US radio stations at Yahoo
The list shows each US radio station listed by
Yahoo, annotated with a genre (country, rock, classical, etc.) and a
number of stars based on their reputation for that genre as computed
by TOPIC.
-
Known authorities
This is a list of known authorities in a few specific topics
(as identified by HITS) and the reputations of those pages
as identified by TOPIC.
-
Annotated search engine results
This shows how TOPIC can be used to annotate the result of a
search engine with the reputation topic of each page, providing
more descriptive results.
Comments? Questions? Please forward them to Davood Rafiei
at drafiei@cs.ualberta.ca.
This page was reconstructed in 2010 with some help from
Internet Archive!