adfa, p. 1, 2011. © Springer-Verlag Berlin Heidelberg 2011 *Corresponding author Heart Sounds Classification with a Fuzzy Neural Network Method with Structure Learning Lijuan Jia 1 , Dandan Song 1* , Linmi Tao 2 , Yao Lu 1 1 School of Computer Science, Beijing Institute of Technology, Beijing 100081, China 2 Department of Computer Science and Technology, Tsinghua University, Beijing 100084, China seemefly2012@gmail.com, sdd@bit.edu.cn, linmi@mail.tsinghua.edu.cn, vis_yl@bit.edu.cn Abstract. Heart sound analysis is a basic method of cardiac evaluation, which contains physiological and pathological information of various parts of the heart and the interactions between them. This paper aims to design a system for ana- lyzing heart sounds including automatic analysis and classification. With the features extracted by wavelet decomposition and Normalized Average Shannon Energy, a novel fuzzy neural network method with structure learning is pro- posed for the heart sound classification. Experiments with real data demonstrat- ed that our approach can correctly classify all the tested heart sounds even for the ones with previous unseen heart diseases. Keywords: Heart sounds, Fuzzy neural network, Structure learning. 1 INTRODUCTION Heart sounds, or heartbeats, are the noises generated by the beating heart and the re- sultant flow of blood through it. Heart sound can be a basic method of cardiac evalua- tion, which contains physiological and pathological information of various parts of the heart and the interactions between them. In healthy adults, there are two normal heart sounds often described as a lub and a dub (or dup), which occur in sequence with each heartbeat. These are the first heart sound (S1) and the second heart sound (S2), produced by the closing of the atrioventricular valves and semilunar valves re- spectively. Fig. 1 shows the heart sound of the apex under normal circumstances. Fig. 1. An example of normal heart sound signals