RDF Knowledge Graph Mining over Data Cubes

Existing efforts in RDF knowledge graph (KG) mining solely addressed fact-level KGs. Many KGs, especially those published by government entities, however contain aggregated data modeled using data cubes. We applied a state-of-the-art KG mining tool on RDF datasets of two government institutions and on interlinked fact-level KGs (Wikidata and YAGO), in multiple settings. Some of the results are meaningful, the main challenges being the low level of support of individual rules and their difficult readability.