Graph matching

Hi,

I don't know whether this question has been answered before but I have found
no reference, yet :)

I want to store a number of instances of a particular ontology in a sort of
library and then be able to search through those instances to retrieve which
instance more closely matches a particular query. I have thought of two
possible solutions:
1. have some index (in RDF/RDFS/OWL or XML) to help in the storing of these
instances in a hierarchy and then when queried, I can make use of some
reasoner to extract the triples and do the matching (between query and
instance). Which matching algorithms are most suitable in this case? Would a
similarity-flooding algorithm be ideal for such situations? Any other
algorithms for graph matching that I can consider?
2. store every instance in an RDF store and then make use of some query
language + reasoner to perform the matching. Only I think that having all
the instances in one large table is not that efficient.

I have checked with Jena and it seems that it is possible to have multiple
models and query these separately. Has anyone used Jena in this way?
Which is the best solution? (1 or 2)
Does anyone know of any other solutions?

Pointers and comments are greatly appreciated.

Regards,

Charlie

Received on Monday, 11 October 2004 18:51:40 UTC