[Python-ideas] Powerset

[Wes Turner]

On Tuesday, October 16, 2018, Greg Ewing <greg.ewing at canterbury.ac.nz>
wrote:

> Wes Turner wrote:
>
>> Is there a name for an iteration of the powerset which is more useful for
>> binary search? I.e. instead of starting with null set, start with the
>> "middle" ( r/2 ).
>>
>
> You'll have to provide more detail about what you want to search
> and how you intend to search it. There isn't a single "middle" to
> the set of powersets, since in general there are many subsets with
> about half the elements of the original set. Also there is no
> obvious ordering to use for bisection.


When searching for combinations of factors which most correlate to the
dependent variable, it doesn't always make sense to start with single
factors; especially when other factors 'cancel out'.

For example, in clinical medicine, differential diagnosis is a matter of
determining what the most likely diagnosis/es is/are; given lots of noise
and one or more differentiating factors.

Testing individual factors first may not be the most efficient because
combinations/permutations are more likely to be highly correlated with
specific diagnoses.

Random search of the powerset and mutation (or a neuralnet) may be faster
anyways. Just wondering whether there's a name for differently ordered
powerset (and Cartesian product) traversals?

Obviously, this is combinatorics and set theory (category theory (HOTT));
here in the itertools library for iterables.


> --
> Greg
> _______________________________________________
> Python-ideas mailing list
> Python-ideas at python.org
> https://mail.python.org/mailman/listinfo/python-ideas
> Code of Conduct: http://python.org/psf/codeofconduct/
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-ideas/attachments/20181016/056b9639/attachment-0001.html>