Research Article

Heart Disease Classification Using Recurrent Neural Networks with ODE-Based Feature Engineering and SMOTE

Abstract

Heart disease remains a major public health concern, creating a need for reliable computational methods that can assist in risk classification using available clinical measurements. In this study heart disease prediction combines ordinary differential equation (ODE) based feature engineering, the Synthetic Minority Oversampling Technique (SMOTE), and Recurrent Neural Network (RNN). The predictive variables include age, gender, total cholesterol (T.CHL), low-density lipoprotein cholesterol (LDL), high density lipoprotein cholesterol (HDL), and triglycerides (TRI). SMOTE were incorporated to address class imbalance. Random forest hyperparameters are optimized using GridSearchCV with stratified five-fold cross-validation and ROC-AUC as the optimization criterion.