Development of a Standardized Real-Time Electrocardiogram (ECG) Dataset for Cardiovascular Disease Detection and Diagnosis Utilizing Advanced Machine Learning Approaches

Abstract Purpose Cardiovascular diseases (CVDs) are still the leading cause of death worldwide, emphasizing the critical need for reliable diagnostic systems. This study aims to create a standardized electrocardiogram (ECG) dataset that can be used to detect and classify six major CVDs using machine learning techniques and investigate feature selection and extraction methods for improved performance. Methods The Department of Cardiology at Sher-i-Kashmir Institute of Medical Sciences provided the dataset, which included 26,600 training and 7,980 validation entries. The Biocare 1210 ECG machine and ECG Data Interpretation 1000 Software were used to extract key parameters, including heart rate, PR interval, QT interval, QRS duration, and blood pressure. Principal Component Analysis (PCA) and correlation analysis were used to select features and reduce dimensionality to improve model efficiency. Machine learning algorithms such as Random Forest, Gradient Boosting, Decision Tree, Naive Bayes, K-Nearest Neighbors (KNN), and Deep Neural Networks (DNN) were tested using accuracy, precision, recall, and F1-score metrics. Results Random Forest had the highest test accuracy (86.44%), followed by Gradient Boosting (82.68%) and Decision Tree (82.61%). Simpler models, such as Naive Bayes and KNN, performed moderately, while DNN and logistic regression achieved lower accuracies (29.95%). PCA and correlation analysis improved the model's interpretability while reducing computational complexity. Conclusion This study emphasizes the significance of algorithm selection and feature engineering in clinical applications. The dataset and findings pave the way for the creation of dependable AI-powered diagnostic tools that strike a balance between accuracy and interpretability, as well as a valuable resource for advancing CVD detection using ECG data.