Consequence−based Datatype Reasoning in EL: Identifying the Tractable Fragments
Despoina Magka
Abstract
The current dissertation suggests a consequence-based reasoning approach for the EL description logic with numerical datatypes. It provides a set of saturation rules which are used in the formulation of a polynomial classification algorithm. Furthermore, it introduces the notion of safety which is a property that numerical datatypes exhibit and guarantees the preservation of polynomiality. The proposed algorithm is proved to be complete only for the safe datatypes. Additionally, the corresponding reasoning problem for non-safe datatypes is proved to be EXPTIME-hard. Apart from that, a classification of specific instances of datatypes is attempted. Finally, the present work, based on the results it produces, suggest a modification of the EL Profile in OWL 2 in order to add datatype features, which are currently available only in OWL 2.
Details
| Affiliation |
Oxford University Computing Laboratory |
| Year |
2009 |
Links
Related pages
|
Activities |