Entity recognition refers to the process of finding words, phrases or non-text patterns that represent a thing such as person, number or date. It typically precedes other forms of document processingng or Intelligent user interfaces. For example, data-detectors scan emails, web pages or other documents to highlight entities such as dates and offer shorcuts to things that you can do with them. Some datatypes such as numbers and date will use some form of pattern matching such as {[regular expressions}}, others, such as placenames, may involve searching for words or n-grams in public or private databases, a process known as named entity recognition. Often simpler entities, may be used as part of higher-level entity detection; for example an email containing date, location and time could be recognised as refering to an event and thus offer the user the opportunity to add it to their calendar.
Used on pages 216, 406
See also named entity recognition