A Deep Learning-Based Framework for Intelligent Heart Disease Prediction Using ANN and CNN

Sanjana Bala J, Deepu K, A PARVEEN, Akshay Govind S, Rakshitha J N

This paper presents a comparative studyof machine learning and deep learning models for heartdisease prediction using clinical datasets. Exploratorydata analysis identified key physiological featuresstrongly associated with cardiovascular risk. ArtificialNeural Network (ANN) and Convolutional Neural Network (CNN) models were developed and evaluatedusing standard performance metrics. Experimentalresults demonstrate that CNN models outperform ANNmodels in terms of accuracy, generalization, androbustness to class imbalance. The CNN achieved anoverall accuracy of approximately 98%, effectivelycapturing complex non-linear patterns in biomedicaldata. These findings highlight the potential of CNNbased models for reliable and automated heart diseaseclassification in clinical decision support systems.