本研究利用HPLC色谱数据的多元信息融合,进行了武夷岩茶的分类识别研究,建立了武夷岩茶不同原料品种、不同产地区域的分类判别模型,模型对训练集样本的判别准确率达到100%,交叉验证准确率达到97%以上,对外部验证集样本的判别准确率达到85%左右。本研究初步证明了利用多元化学指纹图谱判别分析方法对武夷岩茶进行分类识别的可行性。
The method of discriminant classification of Wuyi yan tea based on multiple chemical fingerprint by conjoining two HPLC chromatograms was studied. The results showed that samples from two raw material cultivars or from two districts could be distinguished effectively by using those discriminant method obtained from the investigation. The recognition accuracy for the training sample set was 100%, cross-validating was above 97% and the recognition accuracy for the testing sample set was about 85%. Those result revealed that it was possible to distinguish or verify Wuyi Yan tea by combining the techniques of multiple chemical fingerprint and discriminant classification.
[1] 余虹莺. 武夷岩茶的现状及发展[J]. 技术与市场, 2005(9): 6-7.
[2] 范骁辉, 叶正良, 程翼宇. 基于信息融合的中药多元色谱指纹图谱相似性计算方法[J]. 高等学校化学学报, 2006, 27(1): 26-29.
[3] 王龙星, 肖红斌, 梁鑫淼, 等. 一种评价中药色谱指纹谱相似性的新方法: 向量夹角法[J]. 药学学报, 2002, 37(9): 713-717.
[4] 成浩, 王丽鸳, 周健, 等. 基于化学指纹图谱的扁形茶产地判别分析研究[J]. 茶叶科学, 2008, 28(2): 83-88.
[5] 成浩, 王丽鸳, 周健, 等. 基于化学指纹图谱的绿茶原料品种判别分析[J]. 中国农业科学, 2008, 41(8): 2413-2418.
[6] DANIELE DEL RIO, AMANDA J. STEWART, WILLIAM MULLEN.HPLC-MS Analysis of Phenolic Compounds and Purine Alkaloids in Green and Black Tea[J]. J Agric Food Chem, 2004(52): 2807-2815.