Classification of ECG Signals using Decision Trees and Linear Discriminant Analysis

The heart's crucial function necessitates practical detection tools like Electrocardiograph (ECG) for recording its electrical activity, aiding in understanding heart muscle contractions and identifying abnormalities. Artificial Intelligence, notably Machine Learning, enhances computer performance for cardiac data analysis without explicit programming, surpassing traditional methods like stethoscopes or ECGs. This study compares ECG signal classification using Decision Trees and Linear Discriminant Analysis (LDA) models with patient data. The methodology involves data acquisition, signal processing, feature extraction, and model application. Decision Trees outperform LDA, achieving 98.95% accuracy through K-Fold cross-validation, whereas LDA achieves 88.82%. Evaluation metrics including precision, recall, and F1-Score underscore Decision Trees' superiority, adeptly handling complexity and nonlinearity compared to LDA. These results affirm Decision Trees' accuracy, stability, and overall performance balance, reinforcing their effectiveness in ECG signal classification.