Context and target configurations for mining RDF data

Association rule mining has been widely studied in the context of basket analysis and sale recommendations [1]. In fact, the concept can be applied to any domain with many items or events in which interesting relationships can be inferred from co-occurrence of those items or events in existing subsets (transactions). The increasing amount of Linked Open Data (LOD) in the World Wide Web raises new opportunities and challenges for the data mining community [5]. LOD is often represented in the Resource Description Framework (RDF) data model. In RDF, data is represented by a triple structure consisting of subject, predicate, and object (SPO). Each triple represents a statement/fact. We propose an approach that applies association rule mining at statement level by introducing the concept of mining configurations.