Character 4-grams as a Tool for Semantic Tagging
スポンサーリンク
概要
- 論文の詳細を見る
For a given word that is not in the lexicon it is often possible to guess about its origins by the peculiar character patterns it exhibits (esp. in the case of foreign words or specific terminology). Our system uses a simple, fast method of using character 4-grams trained on various corpora to identify and classify such words.
- 一般社団法人情報処理学会の論文
- 1998-09-17