Open calebweinreb opened 4 months ago
That makes sense. It will take some time, but I will work through each score/metric and check if there is an example of the weights presented in each notebook as well as checking the docstrings to provide an indication. If you notice any others you think are particularly ambiguous, feel free to drop a note here.
Re JOSS review openjournals/joss-reviews#6889
A big strength of
scores
is the ability to incorporate weights into many of the metrics. But how exactly these weights are incorporated is often unclear. Though the application of weights is obvious in come cases (e.g. RMSE), I don't have any intuition for how weights be usedroc_curve_data
, for example. In these ambiguous cases the docstring should be updated to provide a clearer explanation.