Unigram

Part of speech: noun

Definitions

  1. A single word or character taken on its own as one item in a sequence of text, used in statistical language analysis
  2. One isolated element of a body of text, such as a lone word counted separately when studying language patterns
  3. A sequence of length one in computational linguistics, consisting of a single word or token counted without its neighbours

Etymology: "Unigram" is a plain compound: "uni-", from Latin "unus" meaning "one", joined to "-gram", from Greek "gramma", meaning "something written" or "letter". Put together, a unigram is a single written unit, and in its most common modern use it means one item taken alone from a longer sequence. In computational linguistics and statistics, a unigram is a single word or token considered without regard to its neighbours, in contrast to bigrams (pairs) and trigrams (triples). The same pattern of combining a number prefix with "-gram" gives related terms such as "bigram" and "trigram", which is how the word is typically understood in those fields.