目的 对蒲公英Taraxacum mongolicum的颜色、气味和滋味进行数字化表征，并对蒲公英进行产地区分。方法 通过分光测色计、电子鼻和电子舌对30批次蒲公英粉末的颜色、气味和滋味数据进行雷达图、变异系数及主成分分析（principal components analysis，PCA），从中提取对产地区分有意义的特征传感信号指标，对其进行数据融合，做PCA及聚类分析。结果 对分光测色仪传感信号做变异系数分析，色度系统变异系数a^*＞b^*＞L^*，且a^*的变异系数远大于其余2个传感信号值，选择a^*作为色度特征指标；选择电子鼻传感信号前2个主成分进行建模，PC1贡献率73.6%，PC2贡献率10.9%，累积贡献率为84.5%，结合雷达图进行分析，选择对前2个主成分贡献最大且响应最高的W1W及W5S作为影响不同样品的气味特征指标；对电子舌传感信号做雷达图，选择苦味（Bitterness）、咸味（Saltiness）和鲜味（Umami）对不同样品的响应值较高且差异较大的传感器作为滋味特征指标。以苦、咸、鲜、W1W、W5S和a^*特征指标融合数据分析蒲公英整体性状，选择前2个主成分进行建模，累积贡献率为78.6%，能较好地对蒲公英的产地进行区分。聚类分析欧氏距离＝25时，样品划分为1、2两类，欧氏距离为21时，可划分为Ⅰ、Ⅱ、Ⅲ 3类。通过偏最小二乘判别分析（partial least squares discriminant analysis，PLS-DA）筛选得到W5S、a^*和苦味为蒲公英产地区分的差异特征指标。结论 对“色”“气”“味”特征进行PCA和聚类分析可对蒲公英进行产地区分，为其质量评价提供参考。

Objective To discriminate the origin of Taraxacum mongolicum through digitizing characteristics which are color, odor, and taste. Methods Radar plots, coefficient of variation, and principal component analysis (PCA) were performed on the color, odor, and taste data of 30 batches of T. mongolicum powder using spectrophotometer, electronic nose, and electronic tongue. Meaningful feature sensing signal indicators for production area were extracted from the data and fused for PCA and cluster analysis. Results The coefficient of variation analysis of the spectrophotometer sensing signal showed that the coefficient of variation of the chromaticity system was a^*> b^*> L^*, and the coefficient of variation of a^* was much larger than that of the other two sensing signal values. Therefore, a^* was selected as the chromaticity feature marker. Selecting the first two principal components of the electronic nose sensing signal for modeling, with PC1 contributing 73.6%, PC2 contributing 10.9%, cumulative contribution 84.5%. Analyzing the radar chart and selecting W1W and W5S, which have the highest contribution to the first two principal components, as odor characteristic markers that affect different samples; Making a radar image of the electronic tongue sensing signal and selecting sensors with high and significantly different response values for bitterness, saltiness, and umami to different samples as taste characteristic markers. Bitterness, saltiness, umami, W1W, W5S, and a^* characteristic sensing signals were used to analyze the overall traits of T. mongolicum, the first two principal components were selected for modeling, with a cumulative contribution of 78.6%, which can effectively distinguish the origin of T. mongolicum. When the Euclidean distance of clustering analysis = 25, the samples were divided into two categories: 1 and 2. When the Euclidean distance = 21, the samples were divided into three categories: I, II, and III. W5S, a^*, and bitterness were identified as differential characteristic markers for distinguishing T. mongolicum production areas by using partial least squares discriminant analysis (PLS-DA).Conclusion PCA and cluster analysis of "color", "odor", and "taste" characteristics can be used to discriminate the origin of T. mongolicum, which can provide a reference for its quality evaluation.

R286.2

国家中医药管理局科技项目（GZY-KJS-2023-030）；河北省自然科学基金资助项目（H2022423335）；河北省重点研发计划项目（23372502D）；河北省中医药管理局科研计划项目（2025075）