Spotting

Spotting: We call Spotting or Phrase Recognition the task of selecting, from some textual document given as input, phrases that should be annotated by the system. This is closely related to Keyphrase Extraction and Named Entity Recognition, for instance. In Keyphrase Extraction, the system tries to guess the "important" phrases, according to some definition of importance. Meanwhile, in Named Entity Recognition, the system focuses on specific entity types (commonly Person, Location and Organization), and the notion of importance is usually irrelevant. We describe some of these and several other strategies for phrase recognition below.