Automatic Keyword Assignment in English Documents

概要

論文の詳細を見る
Subsequent to the development of the 知的ワ-プロファイリングシステム for Japanese document management, an English version, the Intelligent Document Manager has been developed. In conjunction with this, an automatic keyword assignment algorithm for English documents has been designed. The algorithm performs two major functions a fulltext, syntactic analysis of the target document to extract a list of compact noun phrases;and a statistically-basedevaluation of those noun phrases to find the most suitable keyword candidates. The compact noun phrase extraction locally analyzes parts-of-speech patterns to find short clauses (less than about seven words) of adjectives, nouns and a subset of connectors (conjunctions and prepositions). The keyword candidate evaluation uses word and phrase occurrence frequencies from the document in a straightforward,efficient formula. Smooth system presentation due to two other modules in the package - our proprietary matrix thesaurus and the user interface - allow the system to interact efficiently with the user during keyword registration and document retrieval.
一般社団法人情報処理学会の論文
1990-09-04