Since languages have a defined syntax it would seem logical that a language model that can interpret grammar would have higher accuracy than the unigram model. Why does the unigram model perform better or just as accurately as n-gram or grammar-based models?
No comments:
Post a Comment