Contextus Portal

Brasília, January 2024.

Contextus, developed within the scope of the DFCris project , is an innovation by the Center for Studies, Development, and Innovation in Software (CEDIS) and the Brazilian Institute of Information in Science and Technology (IBICT), funded by the Research Support Foundation of the Federal District (FAP-DF). Under the leadership of professors Sergio Freitas and André Lanna , the platform stands out for applying advanced techniques of Artificial Intelligence, Natural Language Processing, and Machine Learning in the categorization and analysis of scientific information.

Employing sophisticated NLP techniques, Contextus performs meticulous preprocessing of the provided text, covering normalization, tokenization, and lexical filtering stages. Using a specialized dictionary, it transforms words into numerical representations that feed highly advanced ML models, trained with extensive datasets aligned with CAPES research areas. Through deep neural networks and classification algorithms, it identifies the most pertinent research area to the text, creating a direct link between the textual query and the academic spectrum.

This approach not only simplifies the identification of fields of interest for researchers and academics but also enhances the search for specialized knowledge, resorting to quantitative methods for semantic interpretation and thematic classification. The platform, accessible at http://contextus.ibict.br, is a robust and innovative tool that promotes efficient and direct interaction with the academic and scientific universe.