Shao Wenhao, Li Yanjie, Diao Songfeng, Jiang Jingmin, Dong Ruxiang
Research Institute of Subtropical Forestry, Chinese Academy of Forestry, No. 73 Daqiao Road, Fuyang County, 311400, Zhejiang, China.
School of Forestry, University of Canterbury, Private Bag 4800, Christchurch, 8140, New Zealand.
Anal Bioanal Chem. 2017 Jan;409(1):115-120. doi: 10.1007/s00216-016-9944-7. Epub 2016 Oct 28.
The quality of Chinese quince fruit is a significant factor for medicinal materials, influencing the quality of the medicine. However, it is difficult to distinguish different types of Chinese quince fruit. The main objective of this work was to use near-infrared (NIR) spectroscopy, which is a rapid and non-destructive analysis method, to classify the varieties of Chinese quince fruits. Raw spectra in the range of 1000 to 2500 nm were combined with linear discriminant analysis (LDA), quadratic discriminant analysis (QDA), and support vector machines (SVMs) for classification. The first three principal component analysis (PCA) scores were used as input variables to build LDA, QDA, and SVM discriminant models. The results indicate that all three of these methods are effective for distinguishing the different types of Chinese quince fruit. The classification accuracies for LDA, QDA, and SVM are 94, 96, and 98 %, respectively. QDA led to high-level classification accuracy of Chinese quince fruit.
木瓜果实品质是药材的一个重要因素，影响着药品质量。然而，区分不同类型的木瓜果实很困难。这项工作的主要目的是使用近红外（NIR）光谱法（一种快速且无损的分析方法）对木瓜果实品种进行分类。将1000至2500 nm范围内的原始光谱与线性判别分析（LDA）、二次判别分析（QDA）和支持向量机（SVM）相结合进行分类。前三个主成分分析（PCA）得分用作输入变量来构建LDA、QDA和SVM判别模型。结果表明，这三种方法都能有效区分不同类型的木瓜果实。LDA、QDA和SVM的分类准确率分别为94%、96%和98%。QDA实现了木瓜果实的高水平分类准确率。
