Motivation
Our motivation is improvement of Page Ranking Algorithm.PageRanking is connectivity-based quality metric. Such connectivity-based metrics do not require retrieving full page content. They only require retrieving the links on a page. PageRanking is defining the importance of a page by calculating the weighted sum of the back links to it. We introduce personalized PageRank vectors to improve PageRank ranking method. We include the user preferences into calculation of the PageRank. We calculate previously built PageRank vectors to rank the search results based on user preferences. We conduct a user study to find out if our approach provides a better ranking mechanism.
Resources
  • B659 WEB Mining Class home page, [html]
  • Nutch project [html]
  • PageRank algorithm [html]
Progress

Here is a link to the project proposal.[.doc]

Here is the project code.[.tar]

Here is the project write-up.[.pdf]

Here is the link to the running project.[.html]

Here is the link to a paper that we are preparing to submit either CIKM04 or WebKDD04. (NEW!!!) [.pdf]

Here is the link to a WORD document version of our first draft. (NEW!!!) [.doc]

Date: 05/23/04

Contact
Mehmet S. Aktas
CGL Showers Building Suite 224,
Bloomington, Indiana 47408
Email:maktas@cs.indiana.edu

Mehmet A. Nacar
CGL Showers Building Suite 222,
Bloomington, Indiana 47408
Email:mnacar@cs.indiana.edu