Hi Daniel, I constructed a web application that uses option 1, the Jena
option. You can see it at http://www.ontohealth.com.au
go to demonstrations> AI reasoning. The is a link to a diagram of the
application at the bottom of the page. I found the Jena option is fast.Phil
On Tuesday, 23 March 2021, 9:00:33 pm AEDT, Daniel Jeller
<[email protected]> wrote:
Dear all,
I am planning to use Jena/Fuseki as backend for my application and wonder what
the best way is to get data inference. I have a data model (based on an
ontology) where (for instance) there is a “Status” class that can have many
levels of sub-class, like Status->Superior->Abbot or Status->Superior->Prior
and so on. I would like to use this in SPARQL queries where I can ask for
individual instances of the main class as well as any subclass. For this I
would like to use inference based on my Ontology (or better, Ontologies). The
data will change frequently as it will be connected to an online editor where
individual instances can be created or modified.
As far as I can tell I have two options:
* Configure the inference in the software with Jena using ModelFactory,
ReasonerRegistry and so on
Advantages: The data is always up-to-date and based on the current version of
the ontologies exposed via URL
Disadvantages: Slow as the model has to be built for each request (is this
correct?)
* Configure the inference for Fuseki by using a Dataset for the live data,
one for the ontologies (they have to be stored in Fuseki?) and a combined
dataset configured to use inference for the union of both other datasets.
Advantages: faster
Disadvantages: It’s more difficult to update the data or ontologies
Are these assumptions correct? Is there a way to speed up the inference
configured in Jena directly? What would be the preferred way to realize such a
service?
Any advice is greatly appreciated.
Best,
Daniel
--
Mag. Daniel Jeller
Monasterium / Digitisation / IT
ICARUS
--
Spaces Central Station
Gertrude-Fröhlich-Sander Str. 2-4
Tower C, Floor 7-9
A-1100 Vienna
AUSTRIA
Web: http://icar-us.eu<http://icar-us.eu/>
Platforms: http://monasterium.net<http://monasterium.net/>, http//matricula.info
Join the ICARUS friends‘ association:
http://4all.icar-us.eu<http://4all.icar-us.eu/>
