Bigram NLP
noun
Definition: A sequence of two adjacent tokens in text, typically two words or two characters; in language modeling, a bigram model estimates the probability of a token conditioned on the immediately preceding token [???]
Examples in context:
“The features have been extracted using TF/IDF model with unigram and bigram.” [Elkazzaz et al. 2021]
“Unlike morpheme boundary precision, distilling the embedding-based segmentation into a bigram model has almost no effect on Rényi efficiency.” [Libovický et al. 2024]
Synonyms: 2-gram
Related terms: n-gram, bigram model, bigram language model, unigram, trigram