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-discussion] Problem in inferrencing

Timothy Redmond tredmond at stanford.edu
Mon Sep 20 13:33:18 PDT 2010


I suspect that  the core issue is what you mean by

> but what happen not that
>
> still each individual has his own data properties after inferencing
>
> and I adjust preferences of reasoner. I don't know why?
>

The attached ontology seems to be what  you are describing.  In that 
ontology a reasoner will infer that x has data property (age=13) and y 
has name Chris.  In particular in the latest Protege 4.1 if I

   1. classify the ontology
   2. configure the reasoner (Reasoner->Configure->Displayed Inferences)
      to show "Data Property  Assertions"
   3. go to the Individuals tab and select person and then select x or  y


then I will see  the inferences that you are talking about.

-Timothy


On 09/19/2010 07:17 AM, Soha Said wrote:
>
> Thanks so much. you got so closed
>
> I do what you said.
>
> now individual  x has data property (name="Chris")
>
> and individual y has data property (age=13)
>
> after they are same the output must be
>
> individual x: name="Chris" and age="13"
>
> and individual y: name="Chris" and age="13"
>
> but what happen not that
>
> still each individual has his own data properties after inferencing
>
> and I adjust preferences of reasoner. I don't know why?
>
> --------------------------------------------------------
>
> From: Timothy Redmond <tredmond at stanford.edu 
> <mailto:tredmond at stanford.edu>>
>
> Maybe  you are asking whether two individuals that are the same as each
> other should have the same inferred object and data property values?
> The answer is yes.  What's more - this I wasn't clear on - this is
> displayed naturally in the latest Protege.  (I knew  from previous
> discussions that data property values can be difficult for a reasoner
> to compute but I didn't entirely understand why this was so.)
>
> I created the attached ontology.  In the latest Protege 4.1, I selected
> my reasoner (I tried both pellet and HermiT).  I changed the reasoner
> preferences (Reasoner->Configure->Displayed Inferences) to display both
> object property assertions and data property assertions.  Then in the
> individuals tab I selected  the class A and the individual  j.  It then
> showed both the inferred data property and inferred object property that
> was retrieved from  the same individual i.
>
> -Timothy
>
>
> On 09/17/2010 10:25 AM, Timothy Redmond wrote:
> > On 09/17/2010 10:05 AM, Soha Said wrote:
> >> Hello,
> >> I put some rule says that: if person x ,student y , y is copluar to
> >> (some relation in my ontology) x, y is nsubj to (some relation in my
> >> ontology) x then x same as y.
> >> this rule when applied data properties and object properties must be
> >> equaled  of x and y.but what happened object properties are equaled
> >> but data properties are not equaled. why?
> >
> > Again I have trouble understanding your question.  The rule you
> > describe  is as follows:
> >
> >    Person(x)
> >    Student(y)
> >    y copluar x
> >    x nsubj y
> >    ---------------
> >    x = y
> >
> >
> >
> > I am not sure if there are any SWRL constraints  about the conclusion
> > (x=y) but I will ignore that for the moment.  There are not data
> > properties in this rule.  In this rule, x and y must be individuals so
> > the properties, copluar and nsubj,  must be object properties.  I
> > would have guessed that the x and y are the only variables in this
> > rule so this rule only applies to people and students.  It cannot be
> > applied to data properties.
> >
> > -Timothy
> >
> >
> >> Thanks in advance,
> >> Soha Said
> >> 
> --------------------------------------------------------------------------------------------------------------------------------------
> >> Subject: Re: [protege-discussion] Problem in inferrencing
> >> Looking at this, it is hard to tell what you are trying to do.  It
> >> looks like you are asserting that
> >>
> >> all students are people
> >> there is some person who is named Chris.
> >> there is some student, possibly the same as the person above, who is a
> >> graduate.
> >>
> >> Did I miss something?  From this you want to conclude that the
> >> students name is Chris?  How does this follow?
> >>
> >> -Timothy
> >>
> >>
> >> On Sep 15, 2010, at 11:44 PM, Soha Said wrote:
> >>
> >> > Hello,
> >> >
> >> > Thanks for your reply, but I do what you said and after reasoning
> >> > the object properties are equal not data properties. Why?
> >> > Thanks in advance,
> >> > Soha Said.
> >> >
> >> > On Sep 15, 2010, at 9:54 AM, Soha Said wrote:
> >> >
> >> > > Hello,
> >> > > I use prot?g? 4.1 beta. I build an ontology "Student" is 
> subclass of
> >> > > "Person".
> >> > > If I create an instance from each two classes, then properties of
> >> > > person instance inherited to student instance:
> >> > >
> >> > > If "Person_Instance_1":
> >> > > Type: Person.
> >> > > Name: Chris.
> >> > >
> >> > > And, "Student_Instance_1":
> >> > > Type: Student.
> >> > > Adjective: graduate.
> >> > >
> >> > > I want after inferencing, I can deduce that:
> >> > >
> >> > > "Student_Instance_1":
> >> > > Type: Student.
> >> > > Adjective: graduate.
> >> > > Name: Chris.
> >> > >
> >> > >
> >> > > How can I reach this result?
> >> >
> >> > The only way to reach this result is to assert somewhere that
> >> >
> >> >    Person_Instance_1 sameAs Student_Instance_1
> >> >
> >> > You can't get the result you want with inference any other way.  
> After
> >> > all, there can be man instances of Person named "John", "Bill", etc.
> >> > who are differentFrom Student_Instance_1.  So you have to tell the
> >> > system which instances refer to the same individual.
>
>
>
> _______________________________________________
> protege-discussion mailing list
> protege-discussion at lists.stanford.edu
> https://mailman.stanford.edu/mailman/listinfo/protege-discussion
>
> Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03
>    

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.stanford.edu/pipermail/protege-discussion/attachments/20100920/abdfe0a0/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: ChrisIs13.owl
Type: application/rdf+xml
Size: 2736 bytes
Desc: not available
URL: <http://mailman.stanford.edu/pipermail/protege-discussion/attachments/20100920/abdfe0a0/attachment.owl>


More information about the protege-discussion mailing list