[Math] "numerical" mean in distributions?

[Gilles <gi...@harfang.homelinux.org>]

Hello.

In "RealDistribution.java", the Javadoc does not explain why the
method "getNumericalMean()" is named that way, and not simply
"getMean()".

What could the mean be, if not numerical?

Regards,
Gilles


---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org

Re: [Math] "numerical" mean in distributions?

[Phil Steitz <ph...@gmail.com>]

On 1/31/16 4:18 AM, Gilles wrote:
> Hello.
>
> In "RealDistribution.java", the Javadoc does not explain why the
> method "getNumericalMean()" is named that way, and not simply
> "getMean()".
>
> What could the mean be, if not numerical?

If Mikl is still listening, he can chime in with why exactly he
chose to put "numerical" in the characteristic names when these were
introduced in MATH-385; but I suspect it was to make it clear that
what is being returned is not necessarily a closed form, analytical
value, like, e.g. the mean of a Gaussian distribution.  In some
cases, these characteristics are identically NaN (e.g. Cauchy
distribution), in others (e.g. Weibull, Zipf) they are numerical
approximations and in some cases they are exact, computed from the
distribution parameters.  The distinction may be a bit pedantic,
given that we don't make it elsewhere.

Phil
>
> Regards,
> Gilles
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
> For additional commands, e-mail: dev-help@commons.apache.org
>
>



---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org