Slideshow transcript
Slide 1: Reco-Gnizr A bookmark recommendation tool Team - Audumbar Chormale Kishor Datar Uday Acharya
Slide 2: Agenda ● Features Functionality ● System Architecture ● Future Enhancements ● Business Potential ● Demo
Slide 3: Features ● Context aware bookmark recommendation ● Uses user feedback – Rating System ● Embedded in any web page - widget ● Uses Gnizr to create bookmark database
Slide 4: Scenario
Slide 5: System Architecture
Slide 6: How Bookmark Database is built?
Slide 7: Category Identification ● N-gram text classification algorithm ● Algorithm used for – bookmark categorization – Client web page categorization by content scraping ● Wikipedia documents used for training ● Currently supports Arts, Technology, sports
Slide 8: Bookmark Ratings ● User can give feedback on bookmarks recommended by changing rating ● Rating scale : 1 to 5 ● Bookmarks with higher rating are given priority
Slide 9: Web Widget ● Javascript widget that can be embedded in any webpage ● Loads an iFrame in the client page ● Uses JSON data format and AJAX
Slide 10: Future Enhancements ● Customizable widget layout, – Background color scheme – Different Image size ● Fine tuned precomputed ngrams to improve categorization accuracy ● Admin UI for Gnizr: – Generate View Statistics
Slide 11: Summary ● With this service, any user can reap benefits from Gnizr system ● It was a good experience working on – AJAX, JSON – Gnizr API, Webwork framework, Maven – N-gram text classification – Using google code svn
Slide 12: Demo
Slide 13: Thank you




Add a comment on Slide 1
If you have a SlideShare account, login to comment; else you can comment as a guest- Favorites & Groups
Showing 1-50 of 0 (more)