Human Heart Sounds Classification using Ensemble Methods

Abstract

-Efficient diagnosis of cardiac diseases has become increasingly important because cardiac diseases are one of the main causes of decease worldwide. This article presents the research work pertaining to human heart sounds classification using ensemble techniques. In order to validate the classification results, the proposed framework was applied on publicly available standard heart sound dataset. A set of audio features is identified and used for human heart sounds classification. First, using individual classifiers, the sounds classification on the dataset is carried out. The classification results achieved using individual classifiers comes out lower as compared to the existing methods, therefore ensemble technique is applied. This technique proves to be more effective and robust as it increases the overall classification accuracy. The classification accuracies for human heart sound dataset achieved by the proposed methods are higher than the existing solutions.