Predictive Power of RKNN Models in Early Detection of Heart Disease: A Promising Approach for Improved Prognosis

Amidst the aftermath of the COVID-19 pandemic, a surge in cardiovascular complications necessitates timely detection to mitigate risks, particularly in heart disease. A novel approach employing Rule-Based K-Nearest Neighbors (RKNN) aims at early identification of cardiovascular diseases, leveraging K-Means clustering after meticulous data cleaning and preprocessing. Stratified K-fold cross-validation (K=10) addresses target variable imbalances. Our RKNN model, trained on a UCI-based heart disease dataset sourced from Kaggle, demonstrates outstanding performance when compared to alternative models. With a mean accuracy of 94.37%, precision reaching 96.05%, recall at 94.01 % and an F1-score at 94.78^%, our model showcases its efficacy in the early detection of heart disease. This study sheds light on post-COVID-19 cardiovascular challenges and presents a robust computational model poised to significantly contribute to timely interventions in disease related to heart.