A atividade de indexação apoiada por plataformas computacionais

Abstract

Indexing is an activity of a technical and knowledge intensive nature, which enables the information retrieval by users of information retrieval systems. For this reason, this activity must be carried out in a strategic way, always glimpsing the information retrieval, as well as the economical resources for the information units. In order to develop this research, it was motivated to analyze the processes of the manual indexing activity, identifying the main difficulties faced by the index librarian during the execution of these processes and to identify technologies that support the accomplishment of such activity, thus, new possibilities for indexers. In view of the above, the present research had as general objective to analyze the influence of the use of computational platforms in the development of the activities of the index librarian and as specific objectives, it was stipulated: to identify the existing models of the indexing activity and quality indicators that evaluate this activity; to develop a computational platform prototype that supports indexing by applying such a platform in the technical processing sector of an information unit, thereby measuring the interaction of the indexing librarians with the platform, as well as the qualitative measurement of the platform itself based on the indicators of identified in the literature. As for the methodology, it is characterized as a research of an applied nature, with a qualitative approach and with a descriptive character in relation to its objectives. As methodological procedures, experimental and laboratory methods were used. The computational platform developed was applied in the technical processing sector of the Library of the Regional University Foundation of Blumenau (FURB). There, fiftytwo test sessions were performed by the indexing librarian, who first performed the manual indexing of the documents and then performed semi-automatic indexing with the support of the computational platform. Based on the data collected, it was analyzed that the computational platform presented concepts equivalent to the indexer concepts in fifty of the fifty-two test sessions, presenting an overall average of 25% coherence. It was also identified that the platform presented a degree of contribution when it suggested concepts not thought by the indexer during its manual indexing. These concepts were called accepted concepts, and in forty-five of the fifty-two test sessions, there were accepted concepts, showing that 87% of indexations had suggestions of non-equivalent concepts accepted by the indexer. Based on the data analysis and the interview conducted with the indexing librarian, it was concluded that the computational platform had a positive influence on the activities developed by the indexer, presenting a good level of consistency and a high degree of contribution.