论文部分内容阅读
以牵正散及其不同配伍组分的近红外图谱作为聚类分析的对象,在建立混合成分模型的基础上,采用SIMCA聚类分析法快速分类研究牵正散及其不同配伍组分近红外光谱特征。实验中,分别设牵正散全方、缺白附子、缺全蝎和缺僵蚕的4个样品。结果表明:由于牵正散及其不同配伍组分的近红外图谱变异度较大,其光谱聚类的结果较理想,盲样检测的正确率可达90%。因此,如果足够的样本,增加训练集样本数和采样的代表性,加强操作的标准,其准确率将大大提高。所以,近红外光谱与聚类分析相结合的方法可以快速、无损识别复方中药。
The near-infrared spectra of the positive dispersion and its different components were used as the object of cluster analysis. On the basis of the establishment of mixed component model, SIMCA cluster analysis was used to rapidly classify the positive dispersion and its different compatibility components in the near infrared. Spectral characteristics. In the experiment, four samples of Zhengsanfang, Liaobai aconite, L. sinensis, and L. umbellate were set up. The results showed that the result of spectral clustering was better than that of the normal dispersion and its different components. The accuracy of the blind sample detection was up to 90%. Therefore, if there are enough samples, increase the number of training set samples and representative sampling, and strengthen the standard of operation, its accuracy will be greatly improved. Therefore, the combination of near-infrared spectroscopy and cluster analysis can quickly and non-invasively identify compound Chinese herbs.