DistEL: A Distributed EL+ Ontology Classifier

TitleDistEL: A Distributed EL+ Ontology Classifier
Publication TypeConference Papers
Year of Publication2013
AuthorsMutharaju, R, Hitzler, P, Mateti, P
EditorLiebig, T, Fokoue, A
Conference NameProceedings of the 9th International Workshop on Scalable Semantic Web Knowledge Base Systems, co-located with the International Semantic Web Conference (ISWC 2013)
Date Published10/2013
Conference LocationSydney, Australia
KeywordsClassification, DistEL, Distributed Reasoning, EL+, OWL, Scalability

OWL 2 EL ontologies are used to model and reason over data from diverse domains such as biomedicine, geography and road traffic. Data in these domains is increasing at a rate quicker than the increase in main memory and computation power of a single machine. Recent efforts in OWL reasoning algorithms lead to the decrease in classification time from several hours to a few seconds even for large ontologies like SNOMED CT. This is especially true for ontologies in the description logic EL+ (a fragment of the OWL 2 EL profile). Reasoners such as Pellet, Hermit, ELK etc. make an assumption that the ontology would fit in the main memory, which is unreasonable given projected increase in data volumes. Increase in the data volume also necessitates an increase in the computation power. This lead us to the use of a distributed system, so that memory and computation requirements can be spread across machines. We present a distributed system for the classification of EL+ ontologies along with some results on its scalability and performance.

Refereed DesignationRefereed