Graph Summarization for Source Selection of Querying over Linked Open Data

Junxian Li,Wei Wang

DOI: https://doi.org/10.1109/itnec.2017.8284752

2017-01-01

Abstract:How to query Linked Open Data effectively is a challenge due to large amounts of distributed datasets and lacking of overall information about it. Source selection that filters query-relevant datasets is a challenge. We propose a graph structure named 'Predicate Line Graph (PLG)' to summarize dataset by using schema-level information, and then interlink different PLGs to build a global summarization by using schema-level or instance-level mapping information. In experiments, the global view is generated by Linked Open Data sets and then it is evaluated by query sets. The experimental results show that our summarization can select candidate datasets effectively.