You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I would like to plot the residuals between a histogram and a function and could not find a (simple) way to get the uncertainties of the histogram.
I think this functionality is available in poisson_interval where poisson_interval(H.values()) returns the Poissonian uncertainties for the counts in Hist H if I understand correctly.
However it would be useful to have a method implemented for the Hist object that returns this uncertainty directly, something like
H.uncertainties(uncertainty_type="poisson")
And it would be particularly nice if a similarly simple syntax could return uncertainties for the density values returned by H.density().
EDIT: I found the variances() method which seems to return the uncertainties. It would be very useful if the words "uncertainty" or "error" appeared at least once in the documentation in relation to this method.
The text was updated successfully, but these errors were encountered:
I would like to plot the residuals between a histogram and a function and could not find a (simple) way to get the uncertainties of the histogram.
I think this functionality is available in poisson_interval where
poisson_interval(H.values())
returns the Poissonian uncertainties for the counts in HistH
if I understand correctly.However it would be useful to have a method implemented for the Hist object that returns this uncertainty directly, something like
And it would be particularly nice if a similarly simple syntax could return uncertainties for the density values returned by
H.density()
.EDIT: I found the
variances()
method which seems to return the uncertainties. It would be very useful if the words "uncertainty" or "error" appeared at least once in the documentation in relation to this method.The text was updated successfully, but these errors were encountered: