https://repositorio.ifpe.edu.br/xmlui/handle/123456789/489 Thu, 10 Sep 2026 06:48:56 GMT 2026-09-10T06:48:56Z https://repositorio.ifpe.edu.br/xmlui/handle/123456789/2302 Previsão da evasão escolar através da análise de dados e aprendizagem de máquina: um estudo de caso School dropout is a significant global challenge, which compromises the educational, economic and social development of different communities. This work proposes the use of automatic methods based on machine learning to predict and monitor students at risk of dropping out of school, focusing on students at the Instituto Federal de Pernambuco - Campus Jaboatão dos Guararapes. The assessed risk refers to the possibility of classes interrupting their studies before completion, due to academic and sociodemographic factors identified in the data collected. Several algorithms were tested, and Random Forest presented the best performance, surpassing the other models based on the analyzes used. The results show that it is feasible to use predictive models to support decisions and develop policies to reduce school dropout rates. Wed, 08 Oct 2025 00:00:00 GMT https://repositorio.ifpe.edu.br/xmlui/handle/123456789/2302 2025-10-08T00:00:00Z https://repositorio.ifpe.edu.br/xmlui/handle/123456789/2216 TabM para diagnóstico de doença cardíaca: ensembling eficiente de redes neurais em dados tabulares clínicos Cardiovascular diseases represent the main cause of global mortality, being responsible for approximately 19.8 million deaths annually, making early diagnosis fundamental to reducing mortality and improving patients' quality of life. This work investigated the application of the TabM model, a modern deep learning architecture based on dense neural networks (MLPs) and parameter-efficient ensembling that, despite not being the absolute state of the art, demonstrated unique resilience to overfitting among tabular deep learning models and achieved 2nd place ranking in the TabArena benchmark, for diagnosing heart disease using the Cleveland Heart Disease dataset. The objective was to evaluate whether this competitive architecture in general tabular benchmarks would achieve competitive performance with the best results reported in the literature for cardiac diagnosis. The methodology involved selection of relevant clinical features, application of class balancing techniques, hyperparameter optimization, and stratified cross-validation. The results did not fully support the experimental hypotheses, leading to the acceptance of the null hypothesis. The model demonstrated robustness in AUC-ROC, however it fell below the established criteria in accuracy, recall, and F1-score. Beyond the experimental validation, the study delivered a functional Minimum Viable Product (MVP) in Streamlit for interactive clinical inference, a significant differentiator compared to related works that did not provide operational tools. This work represents the first known application of TabM to heart disease diagnosis, establishing a benchmark for future research and a reproducible open-source pipeline. Wed, 25 Feb 2026 00:00:00 GMT https://repositorio.ifpe.edu.br/xmlui/handle/123456789/2216 2026-02-25T00:00:00Z https://repositorio.ifpe.edu.br/xmlui/handle/123456789/2109 Análise da participação feminina nos cursos da área de computação da rede federal This study analyzes female participation in Computing courses at the Federal Education Network of Brazil, based on the microdata provided on the Nilo Peçanha Platform from 2017 to 2023. The analysis was conducted in two stages. Initially, a descriptive quantitative approach was adopted, using Power BI to examine the data from all Computer Science courses in the Federal Network. Next, only the data from the Federal Institute of Pernambuco (IFPE) was considered, in order to conduct a predictive analysis using Machine Learning techniques, with the aim of identifying variables capable of predicting student dropout in the Computing courses of this institution. The results indicate a continuous growth in female participation, reaching 37.64% at IFPE in 2023, a percentage higher than the national average. However, dropout rates are critical among low-income and brown students. The Random Forest algorithm showed the best performance in predicting dropout risk, with an accuracy of 70.32% and a recall of 84.67%. It is concluded that socioeconomic and racial factors have a greater predictive weight on dropout rates than gender alone. Thus, retention policies should prioritize social vulnerability, aiming to ensure the continued presence of women in the field of Computing. Mon, 05 Jan 2026 00:00:00 GMT https://repositorio.ifpe.edu.br/xmlui/handle/123456789/2109 2026-01-05T00:00:00Z https://repositorio.ifpe.edu.br/xmlui/handle/123456789/2049 Guia inteligente: uma ferramenta de acessibilidade para pessoas cegas ou com baixa visão This course conclusion work, proposes the construction of fundamental bases for the creation of an accessibility tool, also known as assistive technology or adaptive technology, for people with visual impairments, whether blind or with low vision, providing for these people a means of facilitating social integration, especially with regard to their academic training, with the objective of helping them during their displacements, in the internal dependencies of the Federal Institute of Education, Science and Technology of Pernambuco, Jaboatão dos Guararapes campus, and, To fulfill this objective, it is necessary to integrate various knowledge, techniques and methods, facilitating accessibility for people with visual impairments, namely: Orientation and mobility, audio description of scenarios, Assistive technologies, working in an integrated way with the most current techniques of Computer Vision and Natural Language Processing, s being these, subfields of artificial intelligence, finally, this work will be built based on the Design Science Research methodology of research and development. Wed, 01 Feb 2023 00:00:00 GMT https://repositorio.ifpe.edu.br/xmlui/handle/123456789/2049 2023-02-01T00:00:00Z