Bewaard in:
| Hoofdauteur: | |
|---|---|
| Formaat: | Recurso digital |
| Taal: | |
| Gepubliceerd in: |
Zenodo
2026
|
| Online toegang: | https://doi.org/10.5281/zenodo.18508081 |
| Tags: |
Voeg label toe
Geen labels, Wees de eerste die dit record labelt!
|
Inhoudsopgave:
- <p>This project provides two knowledge graphs that we created for the two triple store benchmarks: <a href="https://link.springer.com/chapter/10.1007/978-3-031-70365-2_12"><strong>CMPVY(Carcinogenesis, Mutagenesis, Premier League and Vicodi)</strong></a>, and <a href="https://papers.dice-research.org/2022/ISWC_Hashing_the_Hypertrie/iswc2022_hashing_the_hypertrie_public.pdf"><strong>SWDW(Swdf, Watdiv, Dbpedia, Wikidata)</strong></a>. Here are some more details:</p> <h3>1. Preprocessing</h3> <ul> <li> <p>We preprocessed the <a href="https://downloads.dbpedia.org/2016-10/core-i18n/en/"><strong>DBpedia</strong></a> reference graph in <a href="https://papers.dice-research.org/2022/ISWC_Hashing_the_Hypertrie/iswc2022_hashing_the_hypertrie_public.pdf"><strong>SWDW</strong></a> by:</p> <ul> <li> <p>Removing properties of the <code>http://dbpedia.org/property/</code> namespace.</p> </li> <li> <p>Inferring the classes of all entities based on the class hierarchy.</p> </li> </ul> <p> </p> </li> </ul> <h3>2. Knowledge Base Structure</h3> <p>In the first step of our benchmarking framework, we generate a knowledge graph comprising information from the dataset used during the benchmarking process. Our work include two types of data for each query:</p> <p> </p> <ol> <li> <p><strong>Reference knowledge graph(s)</strong><br>Each query has to be executed on a certain RDF graph. We adopt <a href="https://github.com/dice-group/Lemming"><strong>Lemming</strong></a> to add the following RDF graph features to our knowledge graph:</p> <ul> <li> <p>Number of vertices and edges</p> </li> <li> <p>Number of colourless vertices</p> </li> <li> <p>Minimum, maximum, and average in-degree and out-degree</p> </li> <li> <p>Standard deviation of in-degree and out-degree</p> </li> <li> <p>Graph diameter</p> </li> <li> <p>Average clustering coefficient</p> </li> <li> <p>Number of node triangles and edge triangles</p> </li> </ul> </li> <li> <p><strong>SPARQL query</strong><br>We adopt <a href="https://github.com/AKSW/LSQ?tab=readme-ov-file"><strong>LSQ</strong></a> to add the following SPARQL query features to our knowledge graph:</p> <ul> <li> <p>Entities (<code>dqb:hasEntity</code>), properties (<code>dqb:hasProperty</code>) contained in the query, and the <a href="https://www.w3.org/submissions/2005/SUBM-CBD-20050603/">CBD</a> of the entities</p> </li> <li> <p>Type of query</p> </li> <li> <p>The number of triple patterns</p> </li> <li> <p>The number of basic graph patterns</p> </li> <li> <p>The average degree of vertices</p> </li> <li> <p>The median degree of vertices involved in join operations</p> </li> <li> <p>The minimum, maximum, and median number of triple patterns in a basic graph pattern</p> </li> <li> <p>The presence of certain keywords such as <code>FILTER</code>, <code>DISTINCT</code>, and <code>GROUP BY</code></p> </li> </ul> </li> </ol> <pre><br><br></pre>