Wouter,
I welcome your observations. What I've stated and maybe is not well
understood is that the triples I'm willing to process have no schema
information at all, not even rdf:type. So, if many subjects has properties
'brand', 'price' and 'discount' I 'classificate' them as Products. And if
the new 'type' has an ID and I use it to update the IDs of the subjects in
a given statement I can regard a subject of being of a 'type' because its
ID is a product of the ''type' predicate IDs.
Best regards,
Sebastian.
On Sat, Jun 27, 2015 at 4:07 AM, Wouter Beek <w.g.j.beek@vu.nl> wrote:
> Dear Sebastian,
>
> On Fri, Jun 26, 2015 at 8:52 PM, Sebastian Samaruga <cognescent@gmail.com>
> wrote:
>
>> 'arithmetic' inference
>
> I'm not sure what this is.
>
> In your PDF document you state that: "If a set of Subjects share the same
> set of Predicates the resulting set
> could be regarded as an 'inferred' same type."
>
> This definition of "same type" is flawed since according to it `ex:a` in
> [1] and `ex:b` in [2] would have the same type:
>
> [1] ex:a rdf:type ex:Apple .
> [2] ex:b rdf:type ex:Pear .
>
> Would it not be more accurate to describe what you want to do as
> _classification_?
>
> ---
> Best regards!,
> Wouter Beek.
>
> E-mail: w.g.j.beek@vu.nl
> WWW: www.wouterbeek.com
> Tel.: 0647674624
>