CS298 Proposal
Improving Yioop! User Search Data Usage
Tarun Pepira Ramaswamy(tarun.pepira@gmail.com)
Advisor:Dr.Chris Pollett
Committee Members:Dr. Mark Stamp and Dr. Soon Tee Teoh
Abstract:
In the prior work of Vijaya Pamidi[1], Dr.Pollett's previous student wrote a Firefox extension for Yioop which captures the URL's visited and sends these to the Yioop Search Engine. One of the disadvantages of this extension is the lack of user benefit for sending the data to Yioop. This led to non usage of the extension. Also, the relevance of the URL data sent by the extension is computed in an ad hoc way.
The main goal of my project is to add a visualization component to this Firefox extension encouraging the users to use this extension. The visualization will show the user navigation across the visited urls providing user benefit. The extension will also have a preference pane to accept user configurable data like username and password. Currently, the user clicks on the Yioop search result page are not captured. This extension should capture the user clicks on Yioop search result page and send it to the Yioop Search Engine. My project will also implement a theoretically sound way of computing the relevance of these data, based on modern web reputation systems.
CS297 Results
Created a Firefox Extension with preference pane that helps to store user provided data
Implemented Force Based Algorithm in javascript and used canvas to draw a simple graph
Modified Vijaya's code to index the user clicked URL's in the Yioop search result page
Proposed Schedule
Week 1:
01/26/2012 - 02/1/2012 | Write CS298 proposal |
Week 2-3:
02/02/2012 - 02/15/2012 | Work on creating the visualization graph |
Week 4:
02/16/2012 - 02/22/2012 | Deliverable 1 due: Visualization graph of the visited URL's |
Week 5:
02/23/2012 - 03/1/2012 | Study existing indexing system |
Week 6-7:
03/2/2012 - 03/15/2012 | Research and implement a new way to calculate the relevance of the user data |
Week 8:
03/16/2012 - 03/22/2012 | Deliverable 2 due: Indexing the newly added user data |
Week 9-10:
03/23/2012 - 04/05/2012 | Work on the preference page and integrate all the features |
Week 11:
04/06/2012 - 04/12/2012 | Deliverable 3 due:Demo of the complete working solution |
Week 12:
04/13/2012 - 04/19/2012 | Test all the extensions created and write a document on the tests and results |
Week 13:
04/20/2012 - 04/26/2012 | Work on CS298 Report |
Week 14:
04/27/2012 - 05/03/2012 | CS298 Report first draft- Submit to Advisor and Committee |
Week 15:
05/04/2012 - 05/10/2012 | CS298 Report final document- Submit to Advisor and Committee |
Week 16:
04/11/2012 - 05/17/2012 | Defense |
Key Deliverables:
Software
Deliverable 1:Create a visualization graph for the user visited pages
Deliverable 2:Indexing of newly added user data based on modern reputation system
Deliverable 3:Integration of all features: FireFox extension that captures the user clicked data in Yioop search page, a canvas graph showing the user clicked data and the new indexing technique for for these new clicked data
Report
CS298 Report
Project Code and Test results Documentation
Innovations and Challenges
Indexing the new search data requires extensive research on how to calculate the relevance of the new data
It is challenging to create a visualization tool that captures user clicked data and displays in an aesthetically pleasing way
References
1. http://www.cs.sjsu.edu/faculty/pollett/masters/Semesters/Fall10/vijaya/CS 298 Project Report.pdf
2. https://developer.mozilla.org/en/Building_an_Extension
3. [RANDY2010]Building Web Reputation Systems: By Randy Farmer, Bryce Glass, March 2010
4. [AMY2006]Google's PageRank and Beyond: The Science of Search Engine Rankings: by Amy N. Langville and Carl D. Meyer
5. https://developer.mozilla.org/en/SVG
6. https://developer.mozilla.org/en/HTML/Canvas
|