Fig 1.
Example of an alignment between two ontologies (blue and green).
Circles represent classes, arrows represent subclass relationships between classes, dashed lines represent mappings between classes, and dotted lines represent disjointness restrictions between classes. The circles filled in white represent the classes in the core fragments of the ontologies given the alignment.
Fig 2.
Example of two mappings taken from the UMLS-based Alignment between FMA and NCI ontologies, illustrated as blue and green respectively.
Circles represent classes, arrows represent subclass relationships between classes, dashed lines represent mappings between classes, and dotted lines represent disjointness restrictions between classes. The circles filled in white represent the classes in the core fragments of the ontologies given the alignment.
Table 1.
Total number of classes and disjointness axioms of the matched ontologies, the size of the of the original (unrepaired) UMLS-based alignment, and the number of incoherent classes for each OAEI large biomedical track matching problems.
The number of syntactic, cardinality, functional and Object based disjointness axioms relative to the each matching task are shown in parentheses.
Table 2.
Number of classes in AMLR’s core fragments and checksets, and in LogMap's extracted modules for each matching problem in the OAEI Large Biomedical Ontologies track.
The percentage of classes relative to the original ontologies is shown in parentheses.
Table 3.
Core fragments computation times for each OAEI large biomedical track matching problems.
Table 4.
Evaluation of the repairs produced for the UMLS-based alignments from the OAEI 2013 Large Biomedical Ontologies track.
Mod denotes the total number of modifications to the alignment (with removed mappings shown in parentheses when systems also alter mapping relations) and Inc denotes the number of incoherent classes.
Table 5.
Total number of classes and disjointness axioms of the matched ontologies, the size of the of the original (unrepaired) Bioportal alignment, and the number of incoherent classes for each of five Bioportal matched ontologies.
The number of syntactic, cardinality, functional and Object based disjointness axioms relative to the each matching task are shown in parentheses.
Table 6.
Number of classes in AMLR’s core fragments and checksets, and in LogMap's extracted modules for five Bioportal alignments.
The percentage of classes relative to the original ontologies is shown in parentheses.
Table 7.
Core fragments computation times for five Bioportal Alignments.
Table 8.
Evaluation of the repairs produced for the five Bioportal alignments.
Mod denotes the total number of modifications to the alignment (with removed mappings shown in parentheses when systems also alter mapping relations) and Inc denotes the number of incoherent classes.
Table 9.
Evaluation of the repairs produced with the use of filtering.
F-measure and total number of modifications (Mod) of the repaired AML alignments with and without the use of background knowledge (AML-BK and AML respectively) for the tasks of the OAEI 2013 Large Biomedical Ontologies track, both with and without the use of filtering in the repair procedure.