Palavras de títulos e resumos como acesso ao conteúdo do documento: uma análise numérica

Abstract

The identification of a term or a set of terms which may represent a document, is a basic task of access to the content of this document. The parts that constitute, together with the text, an article – author, title, abstract, citations – are alements of access to the content of articles Considering that, the titles of the articles, although they consist of the accessing points to their content, are insuficient for an adequate indexing and retrieval; that the abstracts are more complete structures than the titles; that the KWIC process is most used and has several advantages, it is hypothesized that, starting from a process of analysis and counting of words from the titles and the abstracts of journal articles, it can be demonstrated that the abstract is an instrument used to enrich the titles, giving more access points to the informative content of literature and being able to incorporate to the KWIC process of indexing For this study it was used a sample of 30 articles of scientific journals in the Chemistry area The words included in the titles and in the abstracts of each article were counted and they totalized 3.350 words with respective frequencies A file was formed according to significant and non-significant words. After the counting preocess of titles and abstracts, the words were submitted to a frequency study and the totals to a correlation analysis he results are presented article by article, and by the totals of 30 articles; 7 charts and 30 grafhics are show.The hypothesis was proved. The abstract is about ten times longer than the title considering the number of words.The potential of enrichment is 12 – for each significant word, which is only in the title, there are 12 other significant words which are only in the abstract.