摘要
为探索快速无损分析云芝提取物品级的方法,采集不同产地云芝提取物的近红外漫反射光谱,通过主成分分析(PCA)对样本原始光谱数据进行降维、压缩,并分别结合偏最小二乘判别法(PLS-DA)和反向传输人工神经网络(BP-ANN)建立识别模型。结果表明:采用主成分分析结合偏最小二乘判别法,建模集和验证集的识别正确率分别为100%和84%;采用主成分分析结合反向传输人工神经网络模型,其建模集和验证集的识别正确率均为100%。由此可见,主成分分析结合反向传输人工神经网络模型可以更好地实现对不同品级的云芝提取物进行分类识别。
In order to explore a rapid and nondestructive method to analyze the grade of Coriolus versicolor extract, near infrared spectra of Coriolus versicolor extract from different places was collected, principal component analysis（PCA） was used to reduce the dimension and compress the original spectral data of the sample, and identification models were established by combining partial least squares（PLS-DA） and back propagation-artificial neurl networks（BP-ANN）. The results showed that the identification accuracies of modeling set and verification set are 100% and 84%, respectively, by using principal component analysis and partial least squares discrimination method. While using principal component analysis combined with back propagation-artificial neurl networks, the recognition accuracies of its modeling set and verification set are 100%. Thus, principal component analysis combined with back propagation-artificial neurl networks model could realize the classification and identification of different grades of Coriolus versicolor extract better.
作者
唐敏
宣爽青
张佳凤
陈丽芳
涂红艳
程志钱
TANG Min;XUAN Shuangqing;ZHANG Jiafeng;CHEN Lifang;TU Hongyan;CHENG Zhiqian(Hangzhou Institute for Food and Drug Control,Hangzhou 310022;Department of Food Science and Technology,Zhejiang University of Technology,Hangzhou 310014;Comprehensive Technology Service Center of Huzhou Entry-Exit Inspection Quarantine Bureau,Huzhou 313000)
出处
《食品工业》
CAS
北大核心
2018年第11期196-200,共5页
The Food Industry
基金
杭州市科技项目(20170432B20
20180432B29)
浙江省公益项目(2017C32095)
湖州市科技计划(2016GY02
2017GY15)
关键词
近红外光谱
云芝提取物
主成分分析
品级分类
near infrared spectroscopy
Coriolus versicolor extract
partial least squares
classification of grade
