Name
Abstract | ||
|---|---|---|
We describe the indexing, ranking, and query processing techniques we implemented in order to process a new kind of SPARQL-fulltext queries that were provided in the context of the INEX 2012 Jeopardy task. The INEX 2012 Linked Data track provides a new data collection that aims to com- bine the benefits of both text-oriented and structured retrievals settings in one unified data collection. For the rapid development of a new query engine that could handle this particular combination of XML markup and RDF-style resource/property-pairs, we de- cided to opt for a relational database system as storage back-end, which allows us to index the collection and to retrieve both the SPARQL- and keyword-related conditions of the Jeopardy queries under one common application layer. To process the 90 queries of the benchmark, we keep two main index structures, one of which is based on a recent dump of DBpedia core triples, and another one, which is based on keywords ex- tracted from the INEX Wikipedia-LOD collection. Additionally, our engine comes with a rewriting layer that translates the SPARQL-based query patterns into SQL queries, thus formulating joins over both the DBpedia triples and the keywords extracted from the XML articles. |
Year | Venue | Field |
|---|---|---|
2012 | CLEF (Online Working Notes/Labs/Workshop) | SQL,Joins,XML,Information retrieval,Computer science,Linked data,Search engine indexing,SPARQL,Relational database management system,Database,Markup language |
DocType | Citations | PageRank |
Conference | 1 | 0.40 |
References | Authors | |
10 | 3 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Arunav Mishra | 1 | 45 | 5.73 |
| Sairam Gurajada | 2 | 118 | 7.83 |
| Martin Theobald | 3 | 1474 | 72.06 |