Automatic indexing and ontologies: the consistency of the research chronology and authoring in the scope of Information Science

Abstract

An alternative to minimize the semantic gap in automatic indexing systems is to make use of knowledge organization systems, like ontologies. Given that these two operational concepts are subjects of study in Information Science (IS), one can deem relevant to identify and analyze the continuity, chronological and authorial consistency of studies between ontologies and automatic indexing. We use as methodology an exploratory / descriptive study based on a systematic review. We conclude that a direct relationship (citation and co-authorship) between the articles analyzed is practically non-existent. Regarding exhaustiveness, specificity, precision and recall rates, we concluded that there are similarities among three of the bases used: LISS, LISTA and ISTA. LSD, the fourth base used, was the one that presented a lower performance in the retrieval, leading to low rates of exhaustiveness and specificity of the indexing process.