Carleton University - School of Computer Science Honours Project
Winter 2019
Constructing Steiner Trees for RDF Datasets
Sharath Kunnanath
SCS Honours Project Image
ABSTRACT
The Resource Description Framework represents a unique, flexible way to store data in the form of triples. Triples consist of a subject, predicate and object entity. Triples are connected to form RDF graphs when the subject of a particular triple is the subject or object of another triple, and the same concept applies to the objects. A large volume of RDF graphs (data sets) can be accessed via remote endpoints on the Linked Open Data cloud. To query the LOD cloud for information, a user must use the query language SPARQL. SPARQL is effective and versatile, but fails to return results from a query if the user does not match the structure of the data set being queried. This project seeks to add functionality to Sapphire, a software written to ease the difficulty in querying RDF data, by algorithmically relaxing the structure of a SPARQL query to return appropriate results. This is achieved by extracting appropriate data from a query and using said data to create a Steiner tree. The Steiner tree will consist of a list of triples, which will provide the user the information they seek without needing to know the extensive structure of the RDF data set being queried.