摘要
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 for three different strains of fruit fly wings vibration sound was established. The AR coefficients and the variance of white noise sequence were extracted as the feature, then the sound of the three different strains of fruit fly wings vibration sound was classified by support vector machine（SVM）. The order of AR model was determined by using AIC criterion, and the parameters of AR model were estimated by Burg method, and then the heavy tailed RBF was used as the kernel function in SVM to implement the feature extraction and classification of the different strains of fruit fly wings vibration sound. The experimental results show that the classification accuracy rate of the three strains of fruit fly wings vibration sound is more than 88%.
出处
《山东大学学报（理学版）》
CAS
CSCD
北大核心
2011年第7期83-86,100,共5页
Journal of Shandong University(Natural Science)
基金
国家自然科学基金资助项目(10974130)
中央高校基本科研业务费专项资金资助项目(GK200901006)
关键词
AR模型
支持向量机
果蝇振翅声
AR model
support vector machine
fruit fly wings vibration sound
