next up previous
Next: Instance Retrieval Up: Evaluation Results Previous: Ontology Satisfiability of OWL Beginning: ORE 2012 Evaluation

Entailment

The evaluation results for entailment evaluation are depicted in Table 19 [*]. FaCT++ errors were datatype and instance related. HermiT took more than 1 hour to solve WebOnt-description-logic-208 test. jcel errors were unsupported axioms. It is also worth mentioning the high number of entailment test cases for which FaCT++ fails to give a correct answer w.r.t. the given reference. In future evaluations we intend to further analyse these cases and check the correctness of the current reference.




Table: Entailment evaluation results
  FaCT++ HermiT jcel
ALT, ms 6 37 22
ART, ms 4 77 12
#tests 55 55 16
CORRECT 2 54 4
INCORRECT 45 0 3
NO-REF 0 0 0
EXCEPTION 8 0 9
TIMEOUT 0 1 0

The evaluation results for non entailment evaluation are depicted in Table 20 [*]. The FaCT++ error was instance related. Test case WebOnt-description-logic-209 took more than 1 hour to be computed by HermiT. The parsing related error and unsupported axiom error was observed with jcel.




Table 20: Non entailment evaluation results
  FaCT++ HermiT jcel
ALT, ms 79 68 N/A
ART, ms 15 216 N/A
#tests 7 7 2
CORRECT 6 6 0
INCORRECT 0 0 0
NO-REF 0 0 0
EXCEPTION 1 0 2
TIMEOUT 0 1 0


next up previous
Next: Instance Retrieval Up: Evaluation Results Previous: Ontology Satisfiability of OWL