Heart Sound Clustering Based on Supervised Kohonen Network
In this paper, a new method based on Supervised Kohonen network (SKN) and Mel-frequency cepstrum coefficients (MFCC) is introduced. MFCC of heart sound signal are extracted firstly, and then features are got by calculating every order of MFCC average energy. Finally, SKN is used to identify heart sound. The experimental result shows that this algorithm has a good performance in heart sound clustering, and is of significant practical value.
X. H. Yang et al., "Heart Sound Clustering Based on Supervised Kohonen Network", Applied Mechanics and Materials, Vols. 138-139, pp. 1115-1120, 2012