Issue
It can be hard to understand :
- what each parameters of smoothNMF does
- what is the relationship between parameters
Example :
It does not make sense to put simplex_W = True and simplex_H = True.
Suggested solution
It would be great to build an object/function that inspects the parameters of smoothNMF and warn the user about possible missuses.
non-exhausistive list of possible conflicts
- simplex_W = True, simplex_H = True
- simplex_W = True, fixed_H is not None
- simplex_H = True, fixed_W is not None
- G = None, simplex_W = True
- The strength of lambda_L, mu, epsilon_reg is much lower/higher than the Poisson loss. (to be checked after a few iterations)
Issue
It can be hard to understand :
Example :
It does not make sense to put simplex_W = True and simplex_H = True.
Suggested solution
It would be great to build an object/function that inspects the parameters of smoothNMF and warn the user about possible missuses.
non-exhausistive list of possible conflicts