The document discusses a method for selecting appropriate lexical paraphrases for children using Japanese dictionaries, with a focus on extracting simple words for difficult vocabulary. It outlines the process of acquiring target word candidates and selecting the most suitable paraphrase based on their similarity in meaning. The results indicate that the proposed method effectively identifies more paraphrasable simple words than comparative methods.
Automatic Selection of Predicates for Common Sense Knowledge Expression长冈技术科学大学 自然言语処理研究室
?
The document discusses the automatic selection of predicates for expressing common sense knowledge (CSK) in artificial intelligence. It presents hypotheses regarding the association between nouns and predicates based on their co-occurrence in sentences, aiming to improve knowledge representation in natural language processing. Additionally, it evaluates existing ontologies and the limitations of manually collected resources like the Japanese ConceptNet.
Developing User-friendly and Customizable Text Analyzer长冈技术科学大学 自然言语処理研究室
?
Yuki Miyanishi and Kazuhide Yamamoto. Developing User-friendly and Customizable Text Analyzer. The International Conference on Practical Linguistics of Japanese (ICPLJ8), pp.172-173 (2014.3)
The document discusses a method for selecting appropriate lexical paraphrases for children using Japanese dictionaries, with a focus on extracting simple words for difficult vocabulary. It outlines the process of acquiring target word candidates and selecting the most suitable paraphrase based on their similarity in meaning. The results indicate that the proposed method effectively identifies more paraphrasable simple words than comparative methods.
Automatic Selection of Predicates for Common Sense Knowledge Expression长冈技术科学大学 自然言语処理研究室
?
The document discusses the automatic selection of predicates for expressing common sense knowledge (CSK) in artificial intelligence. It presents hypotheses regarding the association between nouns and predicates based on their co-occurrence in sentences, aiming to improve knowledge representation in natural language processing. Additionally, it evaluates existing ontologies and the limitations of manually collected resources like the Japanese ConceptNet.
Developing User-friendly and Customizable Text Analyzer长冈技术科学大学 自然言语処理研究室
?
Yuki Miyanishi and Kazuhide Yamamoto. Developing User-friendly and Customizable Text Analyzer. The International Conference on Practical Linguistics of Japanese (ICPLJ8), pp.172-173 (2014.3)
14. 1.is-a関係のリンク判定(2/4)
手法2. 固有名詞抽出によるis-a関係でないリンクの判定
親名が固有名詞の場合is-a関係とならない
少年ジャンプ ONE PIECE 新潟県 長岡市
固有名詞 固有名詞
固有名詞は基本的には単語の最小単位なので、それより下位に単語をもつことが
少ない。そのため、親名が固有名詞であるリンクをis-a関係でないリンクとする
形態素解析ツールMeCabを用いる手法
-○県や○市、辞書に登録されている固有名詞を表示
例: 長岡市、遠藤周作
英語Wikipediaを用いる手法
-各形態素の頭文字が大文字なら固有名詞
例:The Beatles
-記事の文中で、単語の頭文字が大文字なら固有名詞
例: Fifteen prime ministers in Japan have studied at University of Tokyo.
14