Mardak / profile

2 stars 7 forks source link

Compute interests from moz_hosts table #79

Closed mzhilyaev closed 10 years ago

mzhilyaev commented 10 years ago

Simplify rule-based interest computation by computing N-best interests using moz_hosts table.

mzhilyaev commented 10 years ago
  1. Client Computes
    • order hosts by frecency
    • start computing edcats interests for top N hosts recording at each N a) the interests b) the frecency of the bottom host
    • compute that for each N between 1 and 10
    • compute that for each 10-th N thereafter
    • return to server for analysis
  2. Client may need to map host names to domains
  3. This data is collected to verify that moz_hosts table produces results comparable to interests computed from full history
mzhilyaev commented 10 years ago

Closed by https://github.com/Mardak/profile/pull/80