Closed cdawei closed 7 years ago
How to evaluate the top-k recommendation with multi-label (multiple ground-truth)?
Suggestions by Lexing: do what imageNet did? give the system credit if any of the prediction is a hit.
Implemented in e1c4e5e
How to evaluate the top-k recommendation with multi-label (multiple ground-truth)?