Heart Disease Prediction Using Machine Learning

The heart disease remains among the leading causes of death worldwide, which emphasizes the importance of early detection to enhance results and avoid complications like heart attacks. Traditional diagnostic methods are often expensive and time-consuming, creating a demand for faster, data-driven solutions. The prediction capabilities of Machine learning technology for heart disease detection have proven effective yet remain limited due to the problem known as class imbalance. computational complexity, and extensive hyperparameter tuning. Existing models like Gradient Boosting, SVM, and Neural Networks stands as their main achievement while they encounter problems with efficiency alongside limited generalization capabilities. This research proposes a new framework which applies the joint use of Synthetic Minority Oversampling Technique (SMOTE) and Naive Bayes Classifier to resolve these problems, balancing computational simplicity with robust predictive power. The combination of SMOTE and Standard Scaler helps overcome class imbalance by creating new samples for minority classes and normalizes features to maintain model stability. Additionally, A novel feature selection method is presented, and the model accuracy is enhanced through this action which identifies the fundamental features that contribute to heart disease prediction. This innovative approach improves interpretability and performance, offering an efficient, scalable, and accessible solution for early heart disease detection, Healthcare systems benefit through improved patient outcomes and decreased healthcare burdens when nurses participate in the design phase of healthcare systems.