Re: Trying to understand Distributions

Robert Kern <[email protected]> Wed, 14 Jul 2021 10:31:27 -0400
Newsgroups gmane.comp.python.scientific.user
Message-ID <CAF6FJiuOKycHLPq+oWYd7zeQkVvBpk81Gk6D1HeFTX-9N_Yt1Q@mail.gmail.com>
On Wed, Jul 14, 2021 at 10:07 AM Keith Sloan <[email protected]> wrote:

> And when I try an add a normal to the mix I have a problem with x axis etc
> and it looks like a normal distribution
> [image: w]
>
> You'll notice that your shape parameter is very high. The Gamma
distribution does asymptotically approach the normal distribution as the
shape parameter increases. Your shape parameter is very large.

One issue is that it looks like your `uminusr` data is truncated
("censored") to omit any values less than 1.8. The fitting procedure in
`fit()` assumes that you are giving it uncensored data. That's driving it
to make the left tail as low as possible and driving the Gamma to look more
normal.


-- 
Robert Kern

_______________________________________________
SciPy-User mailing list
[email protected]
https://mail.python.org/mailman/listinfo/scipy-user
AF0ADDD2-E65A-4B46-90EF-06EDA6DE74A6.jpeg (image/jpeg, 235.9 KB) - not displayed