Optimized Cardiac Disease Prediction Using Cuckoo-Harmony Enhanced Convolutional Neural Networks

Heart disease prediction is a critical challenge in healthcare, and accurate early diagnosis can improve outcomes. A new model is proposed for cardiac disease prediction using swarm intelligence algorithms based on the Cuckoo Search Algorithm and Harmony Search, in combination with Convolutional Neural Networks. The proposed model uses multiple hyperparameter optimization techniques to enhance the accuracy of cardiac disease prediction. A public dataset of 1,025 patient records with 14 attributes was used to test the proposed method. A CNN optimized for cuckoo harmony outperforms more conventional machine learning approaches. The F1-score was 98%, recall was 95%, precision was 100%, and the ROC-AUC score was 100%. The results show that the proposed method was the most accurate model in terms of classification accuracy. This model allows for early detection of cardiac disease in clinical environments.