2

I recently read an article explaining how to get OWL like reasoning on an LPG store in neo4j using GraphScale. But I am very unclear on how to define OWL like rules on an LPG graph in the first place. Can some please explain it to me?
I am basically trying to create a knowledge graph which I would eventually be using in my ML models. Currently, I'm evaluating LPG stores and triple stores but I'm not sure which kind of data modeling and data store to use. What would I lose on if I don't use sophisticated OWL rules?

Kunal Kishore Singh
  • 127
  • 2
  • 10
  • 2
    GraphScale is used for compaction of the OWL knowledgebase, so called *abstraction*. It's nevertheless used in combination with an OWL reasoner like Konclude or HermiT connected via OWL API. In addition, there are bindings for the graph database Neo4J. So in the end, you query Neo4J via Cypher and GraphScale can lead to a performance boost because of a more compact OWL knowledgebase used by the reasoner. So the "rules" are still just OWL axioms (resp. SWRL rules if supported by the reasoner) – UninformedUser Oct 24 '18 at 10:35
  • So I understand them individually but my question is more about how to tie things together. So let's say I have an existing LPG graph, can I use GraphScale somehow to do efficient reasoning on this LPG? If yes, how? – Kunal Kishore Singh Oct 25 '18 at 12:22
  • what do you mean by "how"? I mean, the obvious way would be to use the documentation. But I hope you know that GraphScale isn't available yet? That is mentioned on their [project page](http://derivo.de/en/products/graphscale/) – UninformedUser Oct 26 '18 at 06:55

0 Answers0