基于中红外光谱数据对中药材的鉴别分析
Identification of Chinese Medicinal Materials by Using Mid-Infrared Spectroscopy Data

文章首先选取中红外光谱数据差异较为明显的波数区域，并使用K-means聚类分析法对其进行分析，从而鉴别出药材种类；再次选取同种药材在不同产地的中红外光谱波峰处数值，对这些数据针对产地和需鉴定药材种类进行均值处理，并对相应波段的差值绝对值的方差进行估计，方差最小的则确定为目标药材的产地；最后，采用暴力枚举法对数据进行分析，并利用方差估计进一步进行判断，得出对应种类和产地。
In this paper, the wavenumber regions with obvious differences in mid-infrared spectral data were firstly selected and analyzed by k-means clustering analysis, so as to identify the medicinal materials. The values at the mid-infrared spectrum peaks of the same medicinal materials in different producing areas were selected again, and these data were averaged according to the origin and the species of medicinal materials to be identified. The variance of the absolute difference values of the corresponding bands was estimated, and the one with the smallest variance was determined as the origin of the target medicinal materials. Finally, the violence enumeration method was used to analyze the data, and variance estimation was used to further judge the corresponding species and origin.