Criminal Investigation with Augmented Ontology and Link Prediction

This paper shows that our basic approach to the Knowledge Graph Reasoning Challenge is to create an ontology of motives and means. We also use link prediction based on knowledge graph embedding techniques to solve the case. The ontologies on motives and means are based on hand-crafted in our previous work and are supplemented by an automated ontology created by linking Wikidata. As a result of preliminary experiments, we present an overview of the ontology created at this stage and show how to predict characters with motives for incidents from the knowledge graph of the novel The Adventure of the Speckled Band.