ORIGINAL ARTICLE Adaptive neuro-fuzzy inference system for diagnosis of the heart valve diseases using wavelet transform with entropy Harun Ug ˘uz Received: 17 August 2010 / Accepted: 15 April 2011 / Published online: 26 April 2011 Ó Springer-Verlag London Limited 2011 Abstract Listening via stethoscope is a preferential method, being used by physicians for distinguishing nor- mal and abnormal cardiac systems. On the other hand, listening with stethoscope has a number of constraints. The interpretation of various heart sounds depends on physi- cian’s ability of hearing, experience, and skill. Such limi- tations may be reduced by developing biomedical-based decision support systems. In this study, a biomedical-based decision support system was developed for the classifica- tion of heart sound signals, obtained from 120 subjects with normal, pulmonary, and mitral stenosis heart valve diseases via stethoscope. Developed system comprises of three stages. In the first stage, for feature extraction, obtained heart sound signals were separated to its sub-bands using discrete wavelet transform (DWT). In the second stage, entropy of each sub-band was calculated using Shannon entropy algorithm to reduce the dimensionality of the feature vectors via DWT. In the third stage, the reduced features of three types of heart sound signals were used as input patterns of the adaptive neuro-fuzzy inference system (ANFIS) classifiers. Developed method reached 98.33% classification accuracy, and it was showed that purposed method is effective for detection of heart valve diseases. Keywords Heart sound  Discrete wavelet transform  Adaptive neuro-fuzzy inference system  Entropy 1 Introduction Heart is one of the vital centers for human life. Since the year 1985, deaths from heart diseases have been ranked second worldwide, right after those from brain infarction [1]. Thus, any of disease that may come out in connection with the heart has a critical value. The relations between the volume, the pressure, and flow of the blood in the heart determine opening and closing of cardiac valves. Normal heart sounds occur as closing of the valves. In addition, the sounds, coming from flow of the blood inside the heart and vessels, are com- ponents of the heart sounds [2]. Heart sounds and murmurs come in general from the movements of myocardial walls, opening and closing of valves, as well as from the flow of blood in and out of chambers [3]. The sound emitted by a human heart during a single cardiac cycle consists of two dominant events, known as the first heart sound S1 and second heart sound S2. While S1 comes from closing of mitral and tricuspid valves, S2 comes from closing of aortic and pulmonary valves [4]. For the analysis of heart sounds, and for their naming within the literature as well, heart has been divided into four regions. These are named as mitral, tricuspid, pul- monary, and aortic regions. These regions are not the anatomical locations of the heart valves, but the direction of blood flow through these valves. Comparing the sounds coming from each region with those coming from other regions, troubled region and reason for the related trouble are attempted to be identified [5]. In this study, using heart sounds obtained from mitral and pulmonary regions, mitral stenosis and pulmonary stenosis diseases have been diagnosed. Abnormalities in the structure of the heart are mostly reflected in the heart sounds [6]. Thus, in order to identify H. Ug ˘uz (&) Department of Computer Engineering, Selc ¸uk University, Konya, Turkey e-mail: harun_uguz@selcuk.edu.tr; harun_uguz@hotmail.com 123 Neural Comput & Applic (2012) 21:1617–1628 DOI 10.1007/s00521-011-0610-x