Indexação automática de documentos textuais: critérios essenciais

Abstract

This study aims to evaluate the criteria for automatic indexing in order to develop software that will be responsible by the extraction of terms which represent the textual content of documents. It is studied the ways to realize the indexing process - manual and automatic - and discussed the application of these criteria for the optimization of the first stage of theprocess, that is the subject analysis. The volume of documents published in the current literature shows that great efforts to acquire alternative techniques for indexing are required. The manual process has been considered a slow process. Through the study of technical and scientific articles in the area, it was identified the criteria for automaticindexing that are most used. Then we analyze the degree of satisfaction obtained by researchers through their combination. Among the goals achieved, there are: (1) list of the criteria available in literature, (2) characterization of each criterion, (3) list of the criteria most used. Besides these, it is the main objective of this work, with the obtaining of a set of selected criteria, form a set considered ideal by the researcher.