Cardiovascular Disease Prediction Using ML Based Classification Methods

Now a day’s heart problems are major cause of death, making it crucial to develop effective diagnostic methods. The early detection of cardiovascular abnormalities can enable timely interventions, leading to enhanced patient health condition and minimizing medical expenses. The development of machine learning (ML) techniques has offered promising avenues to revolutionize the landscape of CVD detection. In this research, classification techniques, including Random Forest (RF), SVM and Multiple Logistic Regression are used to diagnose heart diseases. Applying the patient dataset description, the dataset was normalized and missing values were dealt with, after which feature selection was performed to determine the best predictors. All classifiers were tested using cross-validation technique, and performance is measured using precision, recall, AUC-ROC and accuracy. Each classifier displayed significant strengths in terms of Recall, Precision, and F1-Score for both classes: absence and presence of heart disease. The studies show that RF has slightly higher accuracy as compared to others, but SVM has the highest precision value. As such, these classifiers hold the promise of improving the diagnosis of heart diseases.