Lorenz Gaertner
Lorenz Gaertner
@kratsg As you say, I am using PCA to transform the correlated variables to an equal number of uncorrelated ones. For each of the uncorrelated variables, I determine the corresponding...
Mostly done except for `configuration`.
@mreboud I will create a simple example. I think the main reason I have not done it yet, is that with nested sampling this is not very useful for realistic...
Resolved by #909
I think that optional arguments of tasks should become mandatory here. This ensures backwards compatibility when defaults of optional arguments should change.
How would we deal with fixed constrained parameters then? Would you rather correct for this in the `fit()` function directly, or pass `fix_pars` through to `model_utils.unconstrained_parameter_count` as an optional argument?...
What do you think of this fix? ```python def unconstrained_parameter_count(model: pyhf.pdf.Model, fix_pars: Optional[List[bool]] = None,) -> int: """Returns the number of unconstrained, non-constant parameters in a model. The number is...
No worries, let me know if you want me to change anything.
Either is fine, I guess. I would have just checked after the fit is done. Or do you see any complications with this?