Travel-Hunt-Based Deep CNN Classifier: A Nature-Inspired Optimization Model for Heart Disease Prediction

Abstract

Diagnosing the health condition of heart disease patients using deep learning methods has gained more insights into the healthcare sector and when concerned about the modality of diagnosis, parameters corresponding to the functioning of the heart are considered. Even though there are various methods for predicting heart disease, deep learning methods provide adequate accuracy with less computational time. Hence, a novel method called a travel-hunt-DCNN classifier is proposed in this research. The importance of this research depends on the travel-hunt algorithm, which tunes the hyperparameters in the classifier on the poaching and hunting nature. Additionally, the herding-exploring algorithm enhances the feature selection process, providing better prediction results. The execution is undertaken using the heart disease database from the UCI repository using the performance metrics such as accuracy, sensitivity, specificity and F1 measure. For dataset-3 regarding the TPs, the travel-hunt optimization-DCNN classifier method model achieved an accuracy of 96.665% and the sensitivity and specificity values are 99.220% and 94.639%, respectively.