【摘要】 目的:建立大黄真伪的鉴别方法。方法:本文对52个不同品种和不同产地的大黄样品进行了近红外谱图扫描,用支持向量机(SVM)中4种不同的核函数对近红外谱图进行了正品和非正品大黄的鉴别。结果:鉴别正确率均可达98.1%。结论:讨论了影响4个核函数预算的各个参数,表明多形式核函数更适合于本实验样品的鉴定。同时将此实验结果与用 RBF 神经网络鉴别的结果进行了比较,表明 SVM 使用简便,泛化能力强,核函数选择灵活性大,是大黄样本鉴别的简单可靠的方法。更多还原

【Abstract】 To develop a method for identification of rhubarb.Method:Based on near infrared spectra, this paper applies four different kernel functions of support vector machine(SVM)to identification of fifty-two dif- ferent samples of rhubarb.Results:The correct rates of recognition are 98.1%.Conclusion:The parameters affect- ing four different kernel functions prediction are discussed.Polynomial kernel function is the best one for recognition. Compared with the result of radial basis network recognition,the result indicates that the SVM method is characterized by easy use,well generalization,and good selectivity,so it is an easy and reliable method for rhubarb recognition.更多还原