Home |  ENGLISH |  Kontakt |  Impressum |  Anmelden |  KIT

Techreport3034

Aus Aifbportal

Wechseln zu: Navigation, Suche


Multi-objective Linked Data Query Optimization




Published: 2012 November
Institution: Institut AIFB, KIT
Erscheinungsort / Ort: Karlsruhe
Archivierungsnummer:3034

BibTeX
Kurzfassung
With the rapid proliferation of Linked Data on the Web, the topic of Linked Data query processing has recently gained attention. Works in this direction do not assume full availability of SPARQL endpoints. As an alternative to federation over these endpoints, this new querying paradigm follows the Linked Data principles to use only HTTP lookups for accessing and querying Linked Data. Existing works focus on the ranking and pruning of sources behind the query URIs, or on the efficient processing of data after they have been retrieved from sources. However, there exists no systematic approach for query plan optimization, especially the kind that considers both of these problems in a holistic way. Further, observing that result completeness is no longer a strict requirement and that there is an inherent trade-off between completeness, execution cost and other criteria, we propose a multi-objective optimization framework. In experiments on real world Linked Data, Pareto- optimal plans computed by our approach show benefits over suboptimal plans generated by existing solutions.

Download: Media:Tr-km-multi-objective-gla-dtr.pdf



Forschungsgruppe

Wissensmanagement