著者
長尾 真 水谷 幹男 池田 浩之
出版者
一般社団法人情報処理学会
雑誌
情報処理 (ISSN:04478053)
巻号頁・発行日
vol.17, no.2, pp.110-117, 1976-02-15
被引用文献数
26

An automatic method is developed which extracts important words(typical words or possibly key words) from Japanese scientific documents. The idea is to extract the words which appear frequently in some payicular documents, but which appear seldom in other documents. To measure this particularity of word usage we utilized the idea of X^2 test. The experiment is done on a text book of chemistry of the middle school, and "Current Bibliography on Science and Technology (Electrical Engineering)" edited by JICST. More than 120,000 Japanese words are handled and the results are fairly satisfactory.

言及状況

はてなブックマーク (1 users, 1 posts)

ある分野にのみよく現れる語を抽出→カイ二乗検定。一応ブクマ。

収集済み URL リスト