论文部分内容阅读
分别对3个不同品系果蝇的振翅声建立了AR模型,提取AR系数和白噪声序列的方差作为特征,然后用支持向量机(support vector machine,SVM)分类同种内的3个不同品系果蝇的振翅声。使用AIC准则确定AR模型的阶数,用Burg方法估计AR模型的参数,用重尾径向基函数作为支持向量机的核函数,实现对不同品系果蝇振翅声的特征提取和分类。实验结果表明3个品系的果蝇振翅声的分类正确率均达到了88%以上。
The AR model was established for the vibration of wing flies in three different lines, and the variance of AR coefficients and white noise sequences was extracted as the feature. Then, three different strains of the same species were classified by support vector machine (SVM) Flutter of the fruit fly. The AIC criterion is used to determine the order of the AR model, the Burg method is used to estimate the parameters of the AR model, and the heavy tail radial basis function is used as the kernel function of the support vector machine to extract and classify the vibration characteristics of the fruit fly vibration of different strains. The experimental results show that the classification accuracy of fruit fly vibration wings of three strains reached more than 88%.