An MLP trained on GA-selected urinary peptides reports AUC 0.975 for coronary heart disease, but the metrics are compromised by data leakage from SMOTE and feature selection applied before cross-validation.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Peptidomic-Based Prediction Model for Coronary Heart Disease Using a Multilayer Perceptron Neural Network
An MLP trained on GA-selected urinary peptides reports AUC 0.975 for coronary heart disease, but the metrics are compromised by data leakage from SMOTE and feature selection applied before cross-validation.