Search Mailing List Archives


Limit search to: Subject & Body Subject Author
Sort by: Reverse Sort
Limit to: All This Week Last Week This Month Last Month
Select Date Range     through    

[protege-owl] FW: Individual realization

Thomas Russ tar at ISI.EDU
Mon Apr 2 09:55:58 PDT 2007


On Apr 1, 2007, at 5:48 AM, K. Kotis wrote:

>
> Dear all,
>
> Is there a known problem with automatic classification of  
> individuals using the "compute inferred types" command of OWL menu  
> in Protégé 3.2.1 (using Pellet as reasoning engine)?
>
> Pellet automatically classifies new primitive classes inserted in  
> the taxonomy, however it cannot automatically classify (realize)  
> the defined classes' individuals under the new promitive ones. Is  
> this valid?
Assuming that I understand your question properly, then YES, Pellet  
is doing the correct thing.

The whole point of a primitive class is to state that there is  
something else (modeled by the "primitiveness") to the class that is  
not being told to the reasoner.  Since this additional primitive  
trait needs to be present to belong to the class, it would be wrong  
for a reasoner to classify any instances under the primitive concept  
unless they were asserted to belong to that primitive concept (or one  
of its descendents).

You will notice that there is some of the same effect on the  
classification of primitive concepts.  No concepts which are not  
explicit subclasses of the primitive concepts will be classified  
underneath them.  That is the essence of the semantics of being  
primitive.




More information about the protege-owl mailing list