Banca de QUALIFICAÇÃO: KLEISON ROBERTO DE SOUZA SILVA

Uma banca de QUALIFICAÇÃO de MESTRADO foi cadastrada pelo programa.
DISCENTE : KLEISON ROBERTO DE SOUZA SILVA
DATA : 08/10/2026
HORA: 08:00
LOCAL: Ambiente Virtual: meet.google.com/wdv-scev-soz
TÍTULO:

Software For The Automated Collection and Cataloging of Technical and Technological Outputs From UNEMAT'S Professional Master’s Programs.

PALAVRAS-CHAVES:

technical and technological product; professional master's degree;
information governance; natural language processing; automated cataloging..

PÁGINAS: 65
GRANDE ÁREA: Outra
ÁREA: Multidisciplinar
RESUMO:

Brazilian professional graduate programs established the mandatory development of
a Technical and Technological Product (TTP) as a graduation requirement. However,
academic practice has focused predominantly on the dissertation report, keeping
TTPs as appendices or annexes, without autonomous deposit channels or
standardized metadata schemes. This configuration results in low information
visibility and fragmented records, hindering institutional monitoring by Technological
Innovation Centers (NITs), such as the Innovation Agency of the State University of
Mato Grosso (AGINOV/UNEMAT), in the timely identification of assets eligible for
intellectual property protection and in technology transfer. The objective of this
research was to develop software for the automated identification, extraction,
structuring, and cataloging of TTP data contained in dissertations from UNEMAT's
professional master's programs. Methodologically, the study was grounded in Design
Science Research (DSR), guided by the stages of the Design Science Research
Process (DSRP). Problem characterization was supported by a bibliographic survey
in the Scopus database and a documentary audit of institutional repositories of top-
rated graduate programs by CAPES, identifying access barriers, broken hyperlinks,
and the absence of sections dedicated to TTPs. Solution requirements were mapped
from CAPES' official taxonomy of 21 types and evaluative criteria. The computational
artifact to be developed features a modular architecture, encompassing PDF file
ingestion, a Natural Language Processing (NLP) engine supported by heuristic rules,
regular expressions, and language models, a relational database, and a control panel
(dashboard). The system will be demonstrated through the processing of institutional
dissertations and evaluated regarding algorithmic accuracy against a reference
standard, as well as managerial utility through a focus group. The software organizes
the information flow of products, enabling traceability and retrieval of applied assets.

MEMBROS DA BANCA:
Presidente - 89608006 - GEOVANA ALVES DE LIMA FEDATO
Externo ao Programa - 93377004 - EVERTON RICARDO DO NASCIMENTO
Externo ao Programa - 90384008 - FERNANDO SELLERI SILVA
Externo à Instituição - GUILHERME TREZ - Unisinos
Externo à Instituição - JULIO CESAR NARDI - IFES