Machine Learning based Identification of Cardiac Function using PCG Signal

The study focuses on discerning normal and abnormal cardiac function via phonocardiogram (PCG) signals, employing various machine learning models including Support Vector Machine (SVM), K-Nearest Neighbor (KNN), Decision Tree (DT), and Random Forest (RF). Extracting features such as mean, skewness, kurtosis, and symmetry index from PCG signals, Random Forest (RF) emerged as the preferred classification model, attributed to its faster convergence time compared to SVM. Precision, recall, and F1 score were utilized to evaluate classifier performance, demonstrating high accuracy and effectiveness, particularly with Random Forest. The methodology, combining machine learning models with PCG signal feature extraction, proves effective in identifying and classifying cardiac function. Notably, the low frequency to high frequency ratio stands at 4.6329 for normal and 0.3388 for abnormal cardiac functions. This approach holds significant potential as a diagnostic tool for detecting normal and abnormal cardiac functions, crucial for effective management of cardiovascular diseases.