authority file

Terms from Artificial Intelligence: humans at the heart of algorithms

An authority file is used to ensure that named people, places or other things can be connected in a dataset. For example, the name 'Alan Dix' might refer to the author of AI, HCI and statistics textbooks, or Alan Dix the theatre maker. An authrity file would allocate each a unique identifier, say AlanDix_276493 and AlanDix_835672 and then these identifiers would be used in orther data records to ensure it is clear whcih Alan Dix is being fereed to. In pseudonomised data the identifeir is deliberately chosen to be anonymous (unlike those in the example) and the authorty file is either kept securely or deleted after use.