Atilika Concepts and Entities

The Atilika service returns two different types of tag suggestions:

Entities

Entities are tags referencing named entities in the real world, such as persons, places, organisations and so on. Atilika entities are defined by Atilika, and completely in Atilika's control. You cannot change the entity-based tag suggestions made by Atilika. Atilika entities are grouped into the following categories:

• PER (persons such as Nelson Mandela, Donald Trump, etc.)

• LOC (locations such as Africa, Spain, New York, Chelsea)

• ORG (organizations such as companies, charities, political parties and so on)

• TTL (titles such as Mr, Dr, Sir etc.)

• FAC (facilities such as hotels, sports grounds, theatres, universities, shopping malls)

• REL (religions)

• NAT (nationalities such as Irish, Indian, Australian)

Concepts

Concepts are more abstract tags, referencing ideas or topics such as politics, medicine, football, crime and so on. Concepts are not defined by Atilika, they are customer-defined tags. Atilika learns how to apply concepts by example. The process can be started by supplying a large number of ready-tagged documents to Atilika. Atilika learns from this sample what concepts to suggest for future content. If continuously supplied with feedback on how concept tags are used, it will continue learning and improving its suggestions.

CUE Semantic can be configured to make use of entities only, concepts only, or both entities and concepts.