Application of Machine Learning and Deep Learning Techniques for Binary Classification of Cardiovascular Disease Detection

Abstract

Heart complication has been one of the recent health challenges in recent times which has led to a lot of untimely death. Hence, early detection of cardiovascular (CVD) disease plays a major role in preventing untimely death. It has been established that higher performance is achieved in classification problem when bootstrap aggregation algorithm is applied. However, this study presents a hybrid approach to achieve higher performance score and reduced variance by separating the problem into sub-problem. This study applied ensemble learning method of Bagging technique using Decision Tree Classifier as the base model and Feedforward Neural Network (FFNN) on a cardiovascular disease dataset to predict a patient with CVD. The two models were evaluated and compared using metrics such as Accuracy, Precision, Recall, F1-Score and Matthew Correlation Coefficient (MCC). The FFNN model achieved Accuracy of 98%, Precision of 0.974, Recall of 0.991, F1-Score of 0.983 and MCC of 0.958 while the Bagging model achieved Accuracy of 97%, Precision of 0.966, Recall of 0.982, F1-Score of 0.974 and MCC of 0.938. The result shows that the FFNN demonstrated superior predictive performance likely due to its ability to capture complex patterns and adaptively learn from data. The study therefore, suggests the implementation of federated leaning for scalable performance using big data paradigm.