IoT-Based Ensemble Method on PCG Signal Classification to Predict Heart Diseases

Heart disease is the leading cause of death in the world, and early detection of cardiovascular disease is important for maintaining overall health conditions. In the medical system, auscultation of the heart is still an essential process, because it is very easy and inexpensive. An automated system will allow this would be extremely satisfying, and affordable heart medical monitoring for the overall population is useful for identifying potential heart abnormalities at an early stage. Through examining the signals from the phonocardiogram, cardiac diagnosis can be done, and the possibility of irregularities can be identified at an early stage. Hence, the design of the phonocardiogram’s smart and automatic analysis tools is very relevant. The PCG signals are obtained as per the primary collection of the Physionet challenge. This goal to evaluate the PCG signal is clinically “healthy” or “unhealthy” in condition. The key improvement in collaborating methods of time feature and frequency feature extracting from PCG enhances accurate identification of cardiovascular disease. The IoT-based ensemble method has been proposed with time, frequency features, and mfcc features. The classifier algorithms used in this work are K nearest neighbor, SVM, and Ensemble. The Ensemble methods obtained better accuracy and sensitivity.