An embedded, ontology-based auto-completion for SPARQL[zur Übersicht]
Marcel Hillesheim will defend his bachelor thesis about “An embedded, ontology-based auto-completion for SPARQL”. The talk is open for the university audience. Due to the current situation, everybody who wants to attend the talk must register via E-mail to firstname.lastname@example.org until 25th June, so we know who will attend and how many people to expect. See the official statement by university for information how to behave on campus in the current situation: https://www.uni-koblenz-landau.de/de/coronavirus
The semantic web gained traction in recent years. However, programming with RDF data sources is not well-integrated in existing development tools. In particular, writing SPARQL queries is challenging. Writing queries requires knowledge about the data, which in turn has to be gained by querying the data manually. This is tedious due to the flexibility of the data. To simplify the exploration of data and formulation of SPARQL queries, this thesis focuses on auto-completion approaches for SPARQL queries and the integration of those approaches within an IDE. The IDE integration should improve the work flow for developers compared to a standalone implementation. We implement a SPARQL auto-completion for the Intellij IDEA. The implementation utilizes two different approaches for SPARQL auto-completions. The first auto-completion approach presented in literature is based on the data itself. It modifies the SPARQL query where a sug- gestion is needed and executes the modified query in the background. The returned results are then displayed to the user as possible suggestions. The second approach presented in literature relies on an ontology for its suggestions. The approach takes the classes and properties defined in the ontology and uses them to provide suggestions. In addition, we provide some basic error-checking by checking queries for satisfiability using the HermiT reasoner. We also introduce a ranking algorithm for the suggestions. Finally, we evaluate the ranking quality and the performance of both approaches using queries from a DBpedia query log.
26.06.20 - 10:15