期刊名称:Encontros Bibli: revista eletrônica de biblioteconomia e ciência da informação
印刷版ISSN:1518-2924
出版年度:2019
卷号:24
期号:55
页码:1-19
DOI:10.5007/1518-2924.2019.e57927
出版社:Departamento de Ciência da Informação – UFSC
摘要:Objective: this study aims to synthetize and classify the noun phrases selection criteria present in methods for automatic indexing by noun phrases of texts written in Portuguese. Methods: The research methodology has an exploratory nature and bibliographic character, and has the content analysis as procedural method. The bases of the noun phrases selection methodologies are criteria as absolute frequency of occurrence, normalized frequency of occurrence, inverse document frequency, non-occurrence in list of stopwords, and the grammatical structure and level of noun phrases. Conclusions: As for the criteria scope, predominates in quantity those based on the noun phrases characteristics (grammatical structure, level, lexical content), in adoption predominates those based on the document content and the corpus content. Results: The main contribution of this work is the panoramic overview of the noun phrases selection criteria for texts written in the Portuguese idiom.
关键词:Automatic indexing;Noun phrases;Noun phrase selection;Portuguese language;Information retrieval;Indexação automática;Sintagmas nominais;Seleção de sintagmas nominais;Língua portuguesa;Recuperação da informação