REACTIVE: A rule-based framework to process reactivity

Abstract:

Ontologies have been successfully used to model data and knowledge that conceptualize a particular domain. However, in the real-world, resource properties can be affected by events, and ontology formalisms need to be able to represent the conditions that fire these events as well as their consequences. In this paper we define the framework REACTIVE that provides the basis to specify active ontologies and to process reactivity efficiently. In REACTIVE, active ontologies are specified in ActRDFS and queries are expressed in ActSPARQL; these languages respectively extent RDFS and SPARQL to specify data reactive behavior. To efficiently process reactivity, REACTIVE implements an optimization technique named Intersection of Magic Rewritings (IMR), which is able to identify the minimal set of changes of the resource properties that are fired by a set of events in a query. We have conducted an experimental study for processing reactivity on a variety of datasets, and we have observed that IMR is able to reduce by at least a half the reactive processing time w.r.t. non-optimized queries.

Año de publicación:

2009

Keywords:

Fuente:

scopus