Closed
Bug 1129928
Opened 10 years ago
Closed 10 years ago
Analyze pairs of url data to extract related domains
Categories
(Content Services Graveyard :: Tiles: Data Processing, defect)
Content Services Graveyard
Tiles: Data Processing
Tracking
(Not tracked)
RESOLVED
FIXED
People
(Reporter: mzhilyaev, Unassigned)
References
Details
(Whiteboard: [story])
Analyze url-pairs data collected via Bug #1110506
The analysis will include
1. frequency filtering
the match is described here: https://docs.google.com/a/mozilla.com/document/d/1o5DB-OFABV0Ze9ye9ve3gyBs-VHIsaLZtFDG58MQoKg/edit#heading=h.xwjr9eu9xn5c
- high frequency sites are excluded based to random assumption
- site pairs are ordered by how unlikely their observed co-occurrence is
- verify that top pairs are meaningful
- build a cluster of "recommending" site whereby if any cluster-site is in user history, any other cluster site can be recommended
2. We may potentially need to reran telemetry experience to identify impressions coming from a single user
|
||
Updated•10 years ago
|
Status: NEW → RESOLVED
Points: 13 → ---
Closed: 10 years ago
Resolution: --- → FIXED
Whiteboard: .? → [story]
You need to log in
before you can comment on or make changes to this bug.
Description
•