RSVM: A Promising Approach for Early Heart Disease Prediction

The cardiovascular complications have rapidly increased after COVID-19 pandemic leading to various health effects, including heart disease. Irregular blood flow and inflammation harm the heart and vessels, driving cardiovascular problems that require early detection to prevent fatalities. A novel approach using supervised machine learning technique called Rule-based Support Vector Machine (RSVM) has been introduced in this paper, aimed at the early detection of this disease. The proposed rule engine is formed after K-Means clustering. Before applying this computational model, the data was cleaned, preprocessed and outliers were removed. The Stratified K-fold (K=10) cross-validation is applied here due to imbalance of target variables. The UCI based dataset of heart disease from Kaggle are utilized to analyze the model’s efficacy. The effectiveness of the proposed model has been assessed by computing various metrics, including mean accuracy (93.7%), precision (96.93%), recall (91.51%) and F1-score (94.06%) that clearly surpasses alternative models in performance.