Class Satisfiability of OWL DL ontologies

Table 7 depicts the qualitative results for the class satisfiability of OWL DL ontologies, hence it only includes the outcomes for FaCT++ and HermiT. FaCT++ errors were datatypes related, while HermiT errors were Java heap space related.

Table 7: Class satisfiability evaluation results

	FaCT++	HermiTl
CORRECT	96	100
INCORRECT	0	0
NO-REF	0	0
EXCEPTION	6	2
TIMEOUT	0	0

93 test cases were completed in less than 10,000 ms. Table 8 shows the average loading and classification times for these ontologies. Both FaCT++ and HermiT provided similar results for these ontologies, however, there were two test cases where the reasoners had disparate behaviours (see Table 9 ).

Table 8: Class satisfiability OWL DL: Loading and reasoning times

	FaCT++	HermiT
# tests	93	93
ALT, ms	504.5	499.9
ART, ms	555.7	335.7

Table 9: Class satisfiability OWL DL: Reasoning times for hard cases

Ontologies	FaCT++ (ms)	HermiT (ms)
fma-lite	799,207	3,761
DLPOnts-Plans	88	17,413