This content has been marked as final. Show 4 replies
No to 1)
A Virtual Model created with the first, the second model and a rulebase will contain the entailments of both the first and second model.
Right now, there is no easy way to query against something like (M1 + inf_for_m1_only) + M2.
However, you may want to ask yourself the question that why M2 does not need to be included in the inference? Is it going to take much
longer time to run create_entailment if M2 is included?
The first graph contains around 300 nodes and the second one has over 50,000 leaves. The first graph makes vast use of transitive "contains" properties and the second also needs to use this property but not transitively since it only has one depth level. Basically my graph has 100+ times as many leaves as nodes and I am trying to optimize for this situation. One of my requirements is to be able to do a model update in under 2 seconds which is impossible if the both leaves and nodes are in a single graph.
Any ideas as to how I could separate the two models in the SEM_APIS (or Jena framework) or satisfy my requirement for <2 second session-commit in some other way would be greatly appreciated.
Whoa! In the other post you mentioned that Oracle 11.2 can do incremental entailment updates. Could this feature solve my problem? Please send me a link to the manual as to how this can be done (I couldn't find it myself).
Edited by: alexi on Sep 9, 2009 2:19 PM
I will post the link to the 11.2 semantic technologies user manual a bit later.
When you say "do a model update in < 2 seconds," does that time include inference maintenance? Just want to clarify.
Please try the following two links