OOSP: Online Ontology Set Picker

Purpose:

Online Ontology Set Picker (OOSP) allows to select, from major repositories, ontologies that satisfy a user-defined sets of metrics. Its main purpose is allowing ontological tool designers to rapidly build custom benchmarks on which they could test different features. It could also serve for usage studies of different ontology language constructs and for pattern spotting. The web front-end allows to specify a broad range of metrics and delivers benchmarks along with their statistics of metrics, including a graph view.

Screencasts

• Introduction to OOSP
• Introduction to Ontology Search Based on Focus Classes and Categorization Power in OOSP
• Introduction to Ontology Search Based on Similarity in OOSP

Latest news:

• 08-11-17, added two new snapshots for LOV and BioPortal (available for search based on exact metrics, for similarity based search and for search based on lexical tokens), see below for more details about collections
• 12-08-16, ontology search based on free text is added
• 24-06-16, ontology search based on similarity extension added (introduced in 'Ontology Search Based on Similarity in OOSP' screencast, see above)
• 22-06-16, ontology visualization tool recommender integrated with OOSP (also mentioned in introductory screencast to OOSP, see above)

Publications:

Zamazal O., Svátek V.: OOSP: Ontological Benchmarks Made on the Fly. In: 1st International Workshop on Summarizing and Presenting Entities and Ontologies at ESWC 2015. Portoroz, Slovenia. 2015.

Zamazal O., Svátek V.: Ontology Search by Categorization Power. In: 2nd International Workshop on Summarizing and Presenting Entities and Ontologies at ESWC 2016. Anissaras, Greece. 2016.

Zamazal O., Svátek V.: Facilitating Ontological Benchmark Construction Using Similarity Computation and Formal Concept Analysis. In: SEMANTiCS 2016 Posters and Demos track. Accepted.

Zamazal O., Dudáš M., Svátek V.: Augmenting the Ontology Visualization Tool Recommender: Input Pre-Filling and Integration with the OOSP Ontological Benchmark Builder. In: SEMANTiCS 2016 Posters and Demos track. Accepted.

OOSP Architecture:

Ontology Repositories:

Currently, it offers two well-curated repositories: BioPortal and Linked Open Vocabularies (LOV). Further, in 2016 we included OntoFarm and NanJing Vocabulary Repository. We processed them using OWLAPI, thus we only include ontologies parseable by OWLAPI. Further, we could not process those ontologies which were not publicly available, were not available (or one of its imported ontologies) from its hosting server. As a results, we have the following snapshots:

• BioPortal February, 2015 snapshot includes 317 out of 375 ontologies, i.e. 85%.
• LOV February, 2015 snapshot includes 461 out of 475 ontologies, i.e. 97%.
• OntoFarm January, 2016 snapshot includes 16 ontologies, i.e. 100%.
• LOV January, 2016 snapshot includes 509 out of 529 ontologies, i.e. 96%.
• BioPortal January, 2016 snapshot includes 399 out of 501 ontologies, i.e. 80%.
• NanJing Vocabulary Repository, 2016 snapshot includes 1403 out of 1532 ontologies, i.e. 92%.
• NanJing Vocabulary Repository merged, 2016 snapshot includes 225 out of 231 ontologies, i.e. 97%.
• (new) LOV November, 2017 snapshot includes 568 out of 617 ontologies, i.e. 92%.
• (new) BioPortal November, 2017 snapshot includes 494 out of 657 ontologies, i.e. 75%.

In future, we want to take snapshots regularly for these repositories. We are also considering to include other well-curated ontology repositories.

By computing BioPortal and LOV overlap, we disclose 5 shared ontologies: biotop, biopax-level3.owl, qudt, tmo and sio.

Computation of overlap between other pairs of ontologies to be appeared soon.

We also stored imports closures for all ontologies in order to make further analysis of ontologies self-supporting.

Ontology Metrics:

We include various ontology metrics from the following 7 categories:

• Entity metrics include numbers of entities (e.g. classes, instances). Newly there is added "Metrics Based on Focus Classes".
• Axiom metrics include numbers of different axiom types (e.g. subsumptions, equivalencies).
• Class expression types metrics include expressions used for construction of anonymous classes (e.g. existential quantification).
• Taxonomy metrics include characteristics of taxonomy (e.g. a number of top classes, leaf classes, branching degree or maximum taxonomy depth).
• OWL2 profiles and reasoning metrics include profile information along with information about consistency and a number of unsatisfiable classes.
• Annotations metrics include numbers of several selected annotation types (e.g. labels, comments, number of different languages involved in labels annotations).
• Detail metrics include some experimentally designed metrics related to domain/range (e.g. number of anonymous classes as domain definition).

We plan to plug in further ontology metrics to support ontology set selection process.