-
build dataset -> Tristan
-
index dataset for elasticsearch -> Tristan
-
try queries -> Tristan
-
parse the logs -> Titouan
-
compute ranks -> Hani
-
user interface -> Lorenzo
-
Start report -> Lorenzo
-
elastic search
- just do it
-
personalization re-ranking
- simple heuristics (bad but easy)
- P-Click
- L-Profile
- S-Profile
- LS-Profile
- G-Click
- graph based
- contextual features
- simple heuristics (bad but easy)
What we want to do:
- validation
- graph analysis
- report
- Readme
- slides