diagnosing avian newcastle, bronchitis and influenza diseases using heart sound signal and support vector machine

this study represents an intelligence procedure for diagnosis simultaneously avian newcastle disease virus, infection bronchitis virus and influenza using heart sound signal. for this aim, the chickens were divided into four groups. the first group was considered as control samples. the second, third and fourth groups were infected with newcastle disease virus, infection bronchitis and avian influenza, respectively. the time domain signals were transferred to the frequency and time-frequency domain using fast fourier transform and discrete wavelet transform. in data mining stage, 25 statistical features were extracted from three domains and the best features were selected using improved distance evaluation (ide) method. the heart sound signals were classified using multiclass support vector machine and dempster-shafer evidence theory. the total accuracy, specificity and sensitivity of classifiers fusion in diagnosing avian diseases were obtained 81.93, 93.29 and 82.28 percent respectively.