cabral-e/diabetes-classificacao-ml

How cabral-e/diabetes-classificacao-ml works

This repository is a reproducible experiment comparing supervised classifiers for diabetes prediction from BRFSS 2015 health indicators. Its main workflow downloads the dataset, explores and prepares the data, splits it into training, validation, and test sets, compares four models on validation metrics, then evaluates the selected SVM on the held-out test set. The script is a single monolithic experiment, so nodes represent its distinct stages rather than separate modules. Relationships rely on the README and the shown script excerpts; unsampled implementation details are not inferred.

Main entry points: comparacao_algoritmos_diabetes.py Validation comparison

Components 15

Data preparation

Dataset download notebook/comparacao_algoritmos_diabetes.py
Data exploration notebook/comparacao_algoritmos_diabetes.py
Data preparation notebook/comparacao_algoritmos_diabetes.py
Holdout split notebook/comparacao_algoritmos_diabetes.py
Preprocessing notebook/comparacao_algoritmos_diabetes.py
SMOTE balancing notebook/comparacao_algoritmos_diabetes.py

Model comparison

KNN model notebook/comparacao_algoritmos_diabetes.py
SVM model notebook/comparacao_algoritmos_diabetes.py
Decision tree notebook/comparacao_algoritmos_diabetes.py
MLP model notebook/comparacao_algoritmos_diabetes.py

Evaluation and results

Validation comparison notebook/comparacao_algoritmos_diabetes.py
SVM test evaluation notebook/comparacao_algoritmos_diabetes.py
Metrics and plots notebook/comparacao_algoritmos_diabetes.py

Other components

Experiment runner
BRFSS dataset

Connections 19

Experiment runner → Dataset download: runs experiment README.md
Dataset download → BRFSS dataset: downloads README.md
BRFSS dataset → Data exploration: provides records notebook/comparacao_algoritmos_diabetes.py
Data exploration → Data preparation: examines data notebook/comparacao_algoritmos_diabetes.py
Data preparation → Holdout split: supplies features notebook/comparacao_algoritmos_diabetes.py
Holdout split → Preprocessing: provides training data notebook/comparacao_algoritmos_diabetes.py
Preprocessing → SMOTE balancing: transforms training data README.md
SMOTE balancing → KNN model: balances training data README.md
SMOTE balancing → SVM model: balances training data README.md
SMOTE balancing → Decision tree: balances training data README.md
SMOTE balancing → MLP model: balances training data README.md
KNN model → Validation comparison: predicts validation README.md
SVM model → Validation comparison: predicts validation README.md
Decision tree → Validation comparison: predicts validation README.md
MLP model → Validation comparison: predicts validation README.md
Validation comparison → SVM test evaluation: selects SVM README.md
SVM model → SVM test evaluation: predicts test set notebook/comparacao_algoritmos_diabetes.py
Validation comparison → Metrics and plots: reports metrics README.md
SVM test evaluation → Metrics and plots: reports metrics notebook/comparacao_algoritmos_diabetes.py

Drawn by GitDiagram on November 13, 2026 from the repository's file tree, README and source files. Read as Markdown ·