Phonocardiogram Signal Based Better Prediction of Aortic and Mitral Stenosis Heart Diseases Using Machine Learning

Valvular Heart Disease (VHD), is a major cause of mortality and morbidity. Even though echocardiography is the most accurate method of diagnosing VHD, not everyone has access to this equipment. It is not sufficient to use manual cardiac auscultation for VHD screening. Since several machine learning models incorporate heart sounds captured with an electronic stethoscope, the diagnosis of VHD may be more precise. Heart valve disorders such as Mitral Stenosis (MS) and Aortic Stenosis (AS) can be detected from the PCG signal. Our goal was to develop an accurate heart sound classification algorithm for severity of AS and MS diagnosis. The approach suggested combines machine learning methods with conventional features. Features are retrieved from PCG signals using the Discrete Wavelet Transform (DWT), Wavelet Packet Transform (WPT), and Perceptual Wavelet Packet Transform (PWPT) techniques. Neighborhood Component Analysis (NCA) technique is applied to all features and the best features are selected. we employ multiple machine learning approaches, including Support Vector Machine (SVM), Random Forest (RF), K-Nearest Neighbors (KNN), and Decision Tree (DT) algorithms. Hyper-parameter optimization is used to train machine learning models in order to improve classification accuracy. In addition, we use our own built-in device to collect PCG signals from renowned heart specialists in Kolkata that have valvular heart diseases in order to confirm the detection of AS and MS with varying degrees of severity (severe AS, mild AS, mild MS, and moderate MS). We then validate the results using our method. In this analysis, the average accuracy of RF, SVM, DT, and KNN were 99.2%, 99%, 98%, and 98.4%, respectively. These results might potentially be useful in multiclass problems involving the classification of stenosis illnesses. The proposed method shows great potential in providing detection of VHD and save the patients from severe cardiac issue.