Profiles & Ontologies

Profiles & Ontologies

Profiles and Ontologies provide a way within the SAIL LABS Media Mining System (MM-System) to organise concepts and make them available for the advanced search and visualisation capabilities offered by the MM-System. Named-Entities (NE’s) or Profiles – think of them as a set of strings, names or labels referring to an entity in texts – form the basic building blocks for the construction of ontologies. They capture the different ways an entity can be referred to in the real-world and across languages. For example, the current president of the United States can be referred to as “Donald Trump” or also as “The Donald”. In both instances, it’s him being mentioned in a document. Such mentions of entities form the basis for many of the statistics and visualisations in the MM-System. The profile they are connected to can also be used for searching such as profile(“Donald Trump”), yielding all documents mentioning any of the associated labels across all languages and scripts. So even a Russian document mentioning Дональд Трамп will be reported as a search result. The MM-System comes with a pre-fabricated set of profiles covering persons, organisations and locations. These can be extended by end-users with proper permissions and training. Further classes of profiles, such as disasters, diseases or financial entities can easily be created using the profile-editor. Ontologies, linking the profiles together and providing structure for searching and visualisation can be created using the ontology-editor included in the Media Mining Client (MMC).But what can a profile actually stand-for? What role does it play within an ontology? What is an ontology and why and when would you want to use it? And what are the resulting benefits?

We’ll address each of these questions in detail in the next chapters of this white-paper. The information presented will enable you to effectively use these features within the MM-System thus increasing your productivity.

Click the download button to get the full use case.

Download White Paper

Close Menu