A Robust Early Diagnosis Heart Disease Prediction System Using Enhanced Deep Belief Networks

In recent years, cardiovascular disease (CVD) has become the leading cause of death worldwide, posing a significant public health challenge. There are a multitude of specific cardiovascular disorders discussed. Multiple variables can increase an individual's susceptibility to heart disease, thus necessitating the need for precise and expeditious methods to identify and treat the ailment. This research suggests a heart disease prediction system that utilizes the enhanced deep belief network (DBN), a DL technology renowned for its ability to automatically extract hierarchical characteristics from intricate data. In the preprocessing stage, PCA is used to decrease the number of dimensions in the dataset and generate principle components that effectively represent the majority of the variation. The aim of this approach is to accurately assess the likelihood of cardiovascular disease (CVD) in patients by analyzing various medical parameters, including age, blood pressure, and cholesterol levels. The model is trained on the well-known Cleveland heart disease dataset and assessed using accuracy, precision, recall, F1-score, and AUC-ROC. Due to its ability to recognize complex data patterns, the DBN can detect heart disease (HD) early. This system achieves achieves 98.73% of accuracy, 96.98% of precision, 97.42% of recall, 98.19% of specificity, 98.69% of sensitivity.