Recognition of Domain-Specific Terms with D-bigram Model

摘要

There have been many researches on the use of corpora in natural language processing (NLP). In this paper, first we describe d-bigram model, which is a statistical model used in NLP. And then we explain the experiments and their results on the influence of corpus size. We take the unknown term recognition as an application, and examine it with various corpus sizes. Empirical results showed that the corpus size changes the reducing ratio of unknown terms. And also, the result indicated that the difference between closed corpora and open corpora should be considered in term recognition.