基于红外光谱分析法的中药材种类与产地判别

Identification of Species and Origin of Chinese Herbs by Infrared Spectroscopy Analysis

针对2021年全国大学生数学建模竞赛E题要求,分别运用删除吸光度标准差相同的波数数据、取一定数量连续波数的吸光度求平均值两种方法,对赛题所给中药材原始数据进行预处理;调用MATLAB工具箱中的classify函数,选择线性判别模式,对已知种类和产地的中药材数据进行判别,并用留一法交叉验证;取得较高准确率时,再用已知数据判别药材的未知种类和产地。经数据测试,药材的种类和产地判别交叉验证准确率最高分别达100%、98%,表明利用红外光谱分析法判别中药材的种类和产地快速准确有效。

To solve the Problem E in the 2021 National College Students Mathematical Modeling Competition,the given original data of Chinese herbs in the competition are preprocessed by using two methods: deleting the wave number data with the same standard deviation of absorbance and averaging the absorbance of a cer tain number of consecutive wave numbers. Using the classify function in the MATLAB toolkit to select a linear discriminant mode, the data of known species and origin of Chinese herbs are identified, and are cross-verified by using “leave one”method. When a high accuracy is obtained, the known data are used to discriminate the unknown species and origin of herbs. By the data testing, the accuracy of cross-validation of species and origin of herbs reach as high as 100 % and 98 % respectively, indicating that the identification of the species and origin of Chinese herbs by infrared spectroscopy analysis is fast and accurate.