摘要：本研究旨在利用随机森林算法对心音进行分类，为心脏疾病的诊断提供依据。本文结构组织如下: 首先通过电子听诊器采集心音，然后基于小波变换对其进行预处理；其次，基于短时傅立叶变换定义并提取时频域有效宽度以表征第一和第二心音的时频域特征；最后，采用随机森林算法对心音进行分类研究以区分正常和异常心音信号。通过高达93.24%分类精度验证了本系统区分正常与异常心音可行性。因此，本研究可以为医护人员或患者提供一种有效的异常心音鉴别方法。 The study aims as utilizing the random forest algorithm to classify heart sounds for diagnosing heart diseases. This paper is organized as follows: the heart sounds are firstly collected via a electronic stethoscope and preprocessed based on the wavelets transform, and secondly the short-time Fourier transform-based (STFT), the frequency domain features and time domain feature are defined and extracted to characterize the features of the first and the second heart sound in time-frequency domain. Finally, the random forest algorithm is employed to classify normal and abnormal heart sounds. The performance evaluation is validated by the achieved accuracy of 93.24% for distinguishing between normal and abnormal signals. Therefore, this study can pro-vide an efficient way to discriminate abnormal sounds for the medical workers or patients.
