Forecasting Heart Disease Risk Using Revised Machine Learning Models

Heart-disease, often synonymous with cardiac arrest or heart attack, stands as one of the predominant contributors to global mortality in our contemporary world. Globally, heart disease claims the lives of approximately 20 million people each year, making up roughly 32% of all global fatalities. Among these, heart attacks account for 60% of the casualties. Heart attacks are gradually increasing among the younger generation which is most alarming. The global surge in heart disease is particularly pronounced in low- and middle-income countries. Due to inadequate preventive care and heart disease risk factor screening, individuals in these regions often experience early-onset heart disease with suboptimal outcomes. This paper has proposed a Revised Logistic Regression (RLR), Revised Random Forest (RRF), and Revised Gaussian Naïve Bayes (RGNB) algorithms to enhance the accuracy, precision, recall, and f1-score of a proposed model that offers a time-efficient and low-risk method for predicting the risk of heart disease. These revised algorithms provide better results compared to Logistic Regression, Random Forest, and Gaussian naïve Bayes. The accuracy of RLR has reached 94.23% which is 6% higher than the previous Logistic Regression algorithm. The accuracy of RGNB has become 90.38% which is 5% higher than previous Gaussian Naïve Bayes. And, the highest increased algorithm which is RRF has become 96.15% which is 9% higher than the previous Random Forest algorithm. Furthermore, the precision has increased to 97%, the recall has become 96%, and the f1-score has become 96%.