You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
According to eq (17) in the original WSDM paper, p(f|c) is based on |v_{c,f}| * std(f).
However, in librec AoBPR implementation here, p(f|c) is calculated with variance(f) directly. I'm wondering whether it is a bug in AoBPR implementation. Correct me please if I misunderstand something.
The text was updated successfully, but these errors were encountered:
According to eq (17) in the original WSDM paper, p(f|c) is based on |v_{c,f}| * std(f).
However, in librec AoBPR implementation here, p(f|c) is calculated with variance(f) directly. I'm wondering whether it is a bug in AoBPR implementation. Correct me please if I misunderstand something.
The text was updated successfully, but these errors were encountered: