Título: I-AutoML: Uma Abordagem Iterativa e Automatizada para Classificação de Dados Usando um Auto Aprendiz
Autores: José Junior, Anne Magaly de Paula Canuto
Resumo: The growing demand for machine learning (ML) solutions across diverse domains has driven the development of tools that automate complex modeling workflows. This paper introduces “I-AutoM”, an interactive and automated system for supervised classification tasks. Built using Python and Streamlit, the system streamlines the entire ML pipeline—from data preprocessing and model training to evaluation and visualization—without requiring user intervention. It integrates eleven popular classifiers, including Random Forest, XGBoost, LightGBM, and CatBoost, and evaluates their performance using five metrics: accuracy, precision, recall, F1-score, and ROCAUC. Tested on ten real-world datasets from Kaggle, the tool demonstrated its ability to adapt to different data structures, offering accessible, explainable, and efficient model comparison. Results highlight how model performance varies across datasets and emphasize the tool’s value for educational, exploratory, and production environments.
Palavras-chave: component; formatting; style; styling; insert.
Páginas: 7
Código DOI: 10.21528/CBIC2025-1191776
Artigo em PDF: CBIC_2025_paper1191776.pdf
Arquivo BibTeX:
CBIC_2025_1191776.bib