Problem Definition:
- Chrome Web Extension to mine web usage topics + recommend web articles
- Construct a
knowledge graph
of the user based on browsing data - Recommend articles based on the graph e.g. User reads articles about Stephen Curry, Tom Brady, and the system recommends articles on Leo Messi, LeBron James, etc.
Data Set Description:
- No dataset for user web browsing
- Instead, use the Never-Ending Language Learner (NELL) from CMU’s Read the Web project to get a pre-trained ontology.
Process Description:
- Collect browsing data on Google Chrome Extension
- Send url list to Flask Application on EC2 Instance
Construct knowledge graph
- Compute Page Rank
- Return recommendations to browser extension (user)
Images: