Improving Heart Disease Diagnosis Through Supervised Machine Learning Techniques

Abstract

Heart disease is a major risk to public health, accounting for 17.9 million deaths globally annually. Predicting cardiac disease with machine learning (ML) approaches like Linear Regression, Random Forests, Naive Bayes, Decision Tree, Neural Network, XGBoost, AdaBoost, Support Vector Machines, and Catboost models, has shown encouraging results and we use ensemble model from the best model depending on their accuracy and others factors. And we find the best accuracy 97.53% from KNN & RF combined. These algorithms use large medical data to provide personalized risk assessment models for each unique user. As a result, heart disease may have a lessening impact on international health systems and improve methods for risk assessment and treatment. The association between cholesterol and fasting blood sugar levels and heart attacks was shown to be the weakest. The risk factors for heart disease are older adults and men. However, it is important to note that these ML approaches are not foolproof and may have limitations in accurately predicting all cases of heart disease. Additionally, further research and development in this area are necessary to enhance the accuracy and reliability of these personalized risk assessment models. It is crucial for healthcare professionals to continue monitoring and staying updated on the latest advancements in ML technology to effectively prevent and treat heart disease in individuals at risk.