Casablanca: Towards a Dataset-Independent Tool for Semantic Web Annotation

Semantic Web Annotation is the process of mapping named entities that appear in free-text to universal identifiers, typically made available by a Linked Data provider such as DBpedia or Freebase. This is a time consuming process, as it is usually performed by human experts with little or no computational assistance. This paper describes Casablanca, a tool built in order to make Semantic Web Annotation an accessible process for any dataset. Casablanca´s approach is to request a minimal configuration describing a dataset. It then provides unified programming and user interfaces to perform annotation based on this configuration. To perform the necessary disambiguation between named entities, Casablanca relies on existing data available in the targeted dataset.