SPARQL Query Writing with Recommendations Based on Datasets

• Published in: Human Interface and the Management of Information. Information and Knowledge Design and Evaluation pp. 310 - 319
• Main Authors: Gombos, Gergő, Kiss, Attila
• Format: Book Chapter
• Language: English
• Published: Cham Springer International Publishing 2014
• Series: Lecture Notes in Computer Science
• Subjects: Federated system; Linked Data; LOD Cloud; Semantic Web; SPARQL
• ISBN: 3319077309; 9783319077307
• ISSN: 0302-9743; 1611-3349
• DOI: 10.1007/978-3-319-07731-4_32

When we write a SPARQL query, we need to know the structure of the dataset. In the relation databases the tables have a scheme, but the semantic data do not have. Autocompletion function exists in SQL environment, but it does not exist in SPARQL environment. We made a system that can help to write SPARQL query. The system has two features. The first is the prefix recommend. We can write shorter queries if we use prefixes because we do not need to write the long IRIs. The second feature is the predicate-based recommendation based on the type of the variable. If a variable is in the query and it has a type condition, then our system recommends further predicates of this type. Our system needs information about the dataset for the recommendation. We can get these information with simple SPARQL queries. The queries run on a federated system. It is useful because the user does not need any information about the endpoints.