Re: [sbml-distrib] Next steps for the Distributions package

[Maciek J. S. <mac...@gm...>]

Hi Lucian,

we use so called 'code names' to identify distributions and parameters. For
example, there is 'Normal1' which stands for normal distribution with the
parameterisation 'mean' and 'stdev', mean and standard deviation,
accordingly. Those code names are specified only in the spec I'm afraid.
Chapter 5 has overview tables for distributions/parameters and their code
names. See Table 5.3 for the univariates.
Regarding what you called restrictions, a precise parameter/support
definition is contained in the Appendix A.2. See e.g. page 214 for 'Normal1'
definition.

Also, useful to know, apart from the
Distribution(MixtureDistribution)/Parameter elements there are also generic
elements such as LowerTruncationBound/UpperTruncationBound which have
optional 'closed'(default)/'open' attributes.

There are no summary statistics like those in UncertML, we have
concentrated on the distributions indeed. In PharmML the summary statistics (we
used the UncertML ones) are defined within its schema directly.

ProbOnto has, as pointed out by Pierre,
- distribution related quantities such as mean, median, mode and variance
and
- distribution defining functions, such PDF/PMF and wherever available also
the cumulative distribution function, CDF, hazard function, HZ, and
survival function,SF.
All these functions, some very often used in modelling, are available again
via the PharmML schema.

Best, Maciej

PS: the issue with the OLS display was described by Pierre, pls see his
email (bottom line: distributions are instances rather then classes in the
ontology).


On Thu, May 11, 2017 at 7:15 PM, Lucian Smith <luc...@gm...>
wrote:

> On Thu, May 11, 2017 at 1:58 AM, Maciek Jacek Swat <mac...@gm...>
> wrote:
>
>> Lucian, you spotted the missing OWL for the 2.5 version.
>> Indeed, it's still due, hope we get it asap out.
>> For the time being could you work with the 2.0 version?
>> It's on OLS
>> https://www.ebi.ac.uk/ols/ontologies/probonto
>>
>
> Yes, of course!  Just wanted to make sure I wasn't missing something.
>
> I can't seem to use OLS to view the parameterization of any
> distribution--I can only find its name, like:
>
> https://www.ebi.ac.uk/ols/ontologies/probonto/terms?iri=
> http://www.probonto.org/ontology%23PROB_c0000020
>
> Is there a way to find the parameterization of the distributions, what
> those parameters are, and what restrictions they have? (i.e. 'must be >=0',
> etc.)
>
> Also, are there ProbOnto summary statistics?  UncertML had a 'statistics'
> branch with things like 'mode' 'range', 'kurtosis', etc.  Is there a
> similar thing in ProbOnto, or is it focused solely on the distributions
> themselves?
>
> -Lucian
>
> ------------------------------------------------------------
> ------------------
> Check out the vibrant tech community on one of the world's most
> engaging tech sites, Slashdot.org! http://sdm.link/slashdot
> _______________________________________________
> sbml-distrib mailing list
> sbm...@li...
> https://lists.sourceforge.net/lists/listinfo/sbml-distrib
>
>