[foaf-dev] Beyond interests

Bob Ferris zazi at elbklang.net
Sat Sep 11 12:38:07 CEST 2010


Am 10.09.2010 13:01, schrieb Bob Ferris:
> Hi,
>
> by reviewing the General User Model Ontology[1,2] (GUMO), I came around
> the question:
>
> Are there further properties re. user "interest" descriptions in FOAF or
> FOAF sub ontologies (in GUMO they used them as auxiliaries)?
>
> For example "has knowledge" or "has ability". These "interests" can then
> also be weighted somehow with the Weighted Interests Vocabulary[3].
> I think this information is also necessary for having a proper user
> model, e.g. to inference something like e.g. "user type" (re. "computer
> skills"), music listening type (e.g. active or passive; savant,
> enthusiast, causal, indifferent), music knowledge type (education, play
> instrument, music experience).
> Maybe, one can say foaf:interest is the super property  of this category
> and wi:WeightedInterest is the super class of the weighting of this
> category instances. However, I do not really know whether this a good
> decision.
>
> So what do you think about this concern? How would you model this
> descriptions?

I read yesterday in the very interesting paper "A Multi-Dimensional, 
Unified User Model for Cross-System Personalization"[1] about the 
dimension of Cognitive Pattern, which includes user-specific aspects 
like interests, knowledge, preference or abilities. They used there the 
facets area-of-interest, competence (with skill and expertise as sub 
facets) and preference for describing the cognitive characteristics of 
an user.
That's why I would propose to introduce a super property, which 
describes this dimension, so that we can apply the object-oriented 
context reification of foaf:interest/foaf:topic_interest to the whole 
cognitive pattern dimension. However, I'm currently unsure, whether e.g. 
foaf:cognitive_pattern or foaf:cognitive_characteristic will be a good 
naming of this super property. So that we can apply the property chain 
to this super property, e.g.

foaf:cognitive_pattern
       owl:propertyChainAxiom
               (wi:preference wi:topic) .

Furthermore, I would tend to replace the foaf:topic_interest property 
here with the commonly used foaf:interest property (I know their 
difference, but I think people are tending to use this property often in 
the case to address directly a thing rather then a page, which describes 
this thing - so the probably misusing the property with it's current 
range).[I think we discussed this topic still somewhere else ;)]
The second question is then, whether to leave wi:preference as it is or 
to rename this property. Another approach might be a kind of Named 
Graphs here, where the 4th element describes the weighting and the 
"interest" dynamics.

So again:  what do you think about this concern? How would you model 
this descriptions?

Cheers,


Bob



[1] 
http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.94.9261&rep=rep1&type=pdf


More information about the foaf-dev mailing list