ANÁLISE DE TÉCNICAS DE SIMILARIDADE TEXTUAL NO REPOSITÓRIO INSTITUCIONAL DO IF GOIANO (RIIF Goiano)

Abstract

Suggestion algorithms have become essential for simplifying effective access to knowledge. These algorithms are frequently employed in recommendation systems, search engines, and academic content platforms, assisting in the customization of the user experience and the identification of relevant materials. In educational and scientific contexts, their implementation can facilitate access to similar studies, encouraging the dissemination of knowledge and the advancement of research. Facing the difficulty of finding works similar to the desired theme, this paper analyzes and compares text similarity techniques in the Institutional Repository of IF Goiano (RIIF GOIANO), focusing on two approaches: data compression similarity and clustering. The algorithms Damicore and K-Means were selected for the analysis. Data collection was performed using a web crawler, followed by the conversion of PDF documents to text. The results indicate that Damicore demonstrates superior efficiency in a qualitative approach, contributing to the organization and accessibility of data in the RIIF GOIANO.