Hello I can't find the page right now, but Virtuoso, at least the open source version, does not have a comprehensive inference support.
In my PhD Thesis, I've 250MB of N3 files, like 10 million triples. Even though it's not really big, I couldn't get inference quickly done within reasonable memory limits (even with pellet) until I simplified my rules of inference to fit in OWL_micro. Now I answer queries using inference within 5-10 seconds each. If it weren't for my need of rules like "A AND B SubClassOf C" I could use RDFS inference, which is even faster. My advice to you would be to create a simpler ontology to get your rules to fit on OWL_Micro. Good luck! -- diogo patrão On Thu, Aug 1, 2013 at 8:53 AM, Bahador(reza)? OFOGHI <[email protected]>wrote: > THanks, I have been looking at Virtuoso as an alternative since most of my > instances are now in TDB. But thanks for the info anyway. If I cannot find > a clean solution with Jena+Virtuoso, then I might use Pellet. > > > > > ________________________________ > From: Dave Reynolds <[email protected]> > To: [email protected] > Sent: Thursday, 1 August 2013 8:22 PM > Subject: Re: OWL reasoning with Jena > > > On 01/08/13 11:06, Bahador(reza)? OFOGHI wrote: > > Hi, > > > > I have a relatively large OWL file (~50MB) I have created as > OntModel.OWL_MEM. Today I tried to bind it with an OWL reasoner using the > following code: > > > > Reasoner reasoner = ReasonerRegistry.getOWLReasoner(); > > reasoner = reasoner.bindSchema(my50MBModel); > > > > And it took forever for the binding process and it did not even return > before I got disappointed and stopped the process. > > > > I wonder if I am missing anything here? Is there any faster way of > defining the model as an inferece model? What should I expect as average > query time on this model then? > > > > Reasoning time is totally dependent on the nature of your data. A small > ontology can be extremely expensive to reason over. > > As a default for the rule based reasoners use the OWL Micro > configuration. It does most of what the fuller configurations do and is > a lot cheaper. > > I trust you are running in-memory and not trying to reason over a TDB or > SDB model. > > If you need complete DL reasoning or if OWL Micro is too slow then use > Pellet or a commercial solution. > > Dave >
