In this paper,classification and characteristics of the Chinese medicine fingerprint spectrum and HPLC fingerprint were studied.The same kind of drug data sample spectra and chromatographic fingerprints were decomposed into the high frequency part and the low frequency part with wavelet.Four groups of data were classified using the k-nearest neighbor method.Results show that whether it is the original data or the data after wavelet transform, spectral data classification results are significantly better than the chromatographic data.It can also be verified that in both the chromatographic data and spectral data, the accuracy of classification of the original data after wavelet transform has been improved greatly.%对中药光谱指纹图谱与色谱指纹图谱的分类特性进行了研究，针对同一种药品数据样品的光谱图谱与色谱图谱进行小波分解，得到其高频部分和低频部分，使用 k-近邻方法对4组经变换后的数据进行分类。结果表明，无论是原始数据还是经小波变换后的数据，光谱数据分类效果都明显好于色谱数据，同时也可以看出，无论是色谱数据还是光谱数据，原始数据经小波变换后再进行分类，分类正确率都要好于原始数据的分类正确率。
展开▼
