The 'data transformation' has following logic axioms before:
has_specified_input only 'data item'
has_specified_input some 'data item'
has_specified_output only 'data item'
has_specified_output some 'data item'
Now 'data transformation' only specify the input in axioms as:
has_specified_input only 'data item'
The logic axioms removed may due to slowing down the reasoning. I performed the test in Protege using Hermit reasoner including:
Pre-computing inferences:
class hierarchy
object property hierarchy
class assertions
has_specified_input only 'data item'
Ontologies processed in 349574 ms by HermiT
has_specified_input some 'data item'
has_specified_output some 'data item'
Ontologies processed in 379260 ms by HermiT
has_specified_input only 'data item'
has_specified_input some 'data item'
has_specified_output only 'data item'
has_specified_output some 'data item'
Ontologies processed in 367654 ms by HermiT
If we added the logic axioms back, it will take about 20 seconds more. Considering reasoning is taking about 350 seconds (~6 min) now, I think 20 secs more is acceptable. So, I'd prefer to add them back.
The 'data transformation' has following logic axioms before:
Now 'data transformation' only specify the input in axioms as:
has_specified_input only 'data item'
The logic axioms removed may due to slowing down the reasoning. I performed the test in Protege using Hermit reasoner including: Pre-computing inferences:
has_specified_input only 'data item'
Ontologies processed in 349574 ms by HermiTOntologies processed in 379260 ms by HermiT
Ontologies processed in 367654 ms by HermiT
If we added the logic axioms back, it will take about 20 seconds more. Considering reasoning is taking about 350 seconds (~6 min) now, I think 20 secs more is acceptable. So, I'd prefer to add them back.