RECAP: Building Relatedness Explanations on the Web

We describe RECAP, a tool that, given a pair of entities defined in some Knowledge Graph (KG), builds an explanation, that is, a graph (of manageable size) reflecting their relatedness. Explanations enable to discover new knowledge and browse toward other entities of interest. We discuss different kinds of explanations based on information theory and diversity. The KG-agnostic approach adopted by RECAP, which retrieves the necessary information via SPARQL queries, makes it readily usable on a variety of KGs.