基于IAFSA-SVM的心音信号识别研究

Research on Heart Sound Signal Recognition Based on IAFSA-SVM

为提高心音信号识别的准确率,针对传统的支持向量机(SVM)在寻找最优核函数参数和惩罚因子时存在的优化问题,提出一种改进的人工鱼群算法(IAFSA)优化SVM的心音信号分类算法(IAFSA-SVM)。首先将采集到的含噪心音信号利用改进的小波阈值进行降噪处理,并进行时频域分析提取出特征值;然后采用改进的人工鱼群算法寻找最优SVM参数,并输入到SVM识别模型中进行心音信号的识别。通过仿真证实IAFSA-SVM算法相对于传统的SVM模型和粒子群优化的SVM模型提高了心音信号识别的准确率,为心脏病的诊断提供了新方法。

In order to improve the accuracy of heart sound signal recognition,in view of the optimization problems of traditional support vector machine(SVM)in finding the optimal kernel function parameters and penalty factors,an improved artificial fish school algorithm(IAFSA)is proposed to optimize the heart sounds of SVM Signal classification algorithm(IAFSA-SVM).First,the collected noisy heart sound signals were denoised using an improved wavelet threshold,and time-frequency domain analysis was performed to extract the characteristic values;Then the improved artificial fish school algorithm was used to find the optimal SVM parameters and input to the SVM recognition model recognition of heart sound signals in.the process.The simulation experiment proves that the IAFSA-SVM algorithm improves the accuracy of heart sound signal recognition compared with the traditional SVM model and the particle swarm optimization SVM model,and provides a new method for the diagnosis of heart disease.