Confidence intervals

[aulemahal]

Pull Request Checklist:

• This PR addresses an already opened issue (for bug fixes / features)
  • This PR fixes Confidence intervals on statistical functions #997
• Tests for the changes have been added (for bug fixes / features)
  • (If applicable) Documentation has been added / updated (for bug fixes / features)
• CHANGES.rst has been updated (with summary of main changes)
  • Link to issue (:issue:number) and pull request (:pull:number) has been added

What kind of change does this PR introduce?

• New confidence argument on stats function that perform a fit.
• Enable interval as a dist method in dist_method.

Does this PR introduce a breaking change?

No, the default behaviour is unchanged.

Other information:

from xclim.testing import open_dataset
import xclim as xc

ds = open_dataset('sdba/CanESM2_1950-2100.nc')
params, confint = xc.indices.stats.fit(ds.pr, 'gamma', confidence=0.95)

[huard]

Hum, I don't think this is exactly what we want. The scipy interval function only returns the bounds a, b such that, with the given parameters, confidence % of samples will be within [a,b]. This is not an indication of the quality of fit.

I'll do some research to clarify the request here.

[huard]

Suggested change

	If set, the interval for the corresponding confidence is returned along the parameters.
	If set, the interval for the corresponding confidence is returned along with the parameters.

[Zeitsperre]

False track. Closing.