Open huard opened 2 years ago
@huard This looks not so hard to implement, but I'm not sure where it should live.
Would it be :
out, confidence_interval = xc.generic.frequency_analysis(da, confidence=0.95)
where
def frequency_analysis(*args):
params = fit(da, dist)
out = parametric_quantile(params, q)
conf_int = dist.interval(0.95, params)
return out, conf_int
Like, is it an optional output for every indicator that computes a fit
?
So the link to the interval
method was a false lead. Sorry for this.
Discussion suggests that this should be contributed upstream to scipy
(or lmoments3
?)
Description
Return confidence intervals (as does R) for statistical estimates.
https://docs.scipy.org/doc/scipy/reference/generated/scipy.stats.rv_continuous.interval.html