Closed alaws-USGS closed 12 months ago
I made a quickie sample notebook to show how the metrics might be used.
Two thoughts:
bias()
function: the docstring says it returns a float. The actual return value is a vector (same len as obs and est). n
on your MAE calc. Can be derived internally by calling len
on the observed vector. -- even better, just let numpy
calc the mean instead of doing it yourself with sum
and divide
: return np.abs(obs - est).mean()
@gzt5142 I updated the mae
function to calculate n internally. However, I wanted to update the method from .mean()
to np.mean()
as this matches up with the array API standard, which numpy is moving towards (from my SciPy23 notes). Or would best practice be to wait to update to that once its fully integrated?
Closes #330