摘要
目的基于不同产地茶叶中微量元素差异应用化学计量学工具构建云南不同产地普洱茶产地溯源模型。方法采用电感耦合等离子体质谱对两个主产区的普洱茶中21种微量元素和15种稀土元素进行检测,应用化学计量学工具对数据进行筛选、主成分分析和产地鉴别模型的构建,与常用Fisher判别方法(Fisher linear discriminant analysis,FLD)比较,对构建模型所采用的决策树算法C5.0(decision tree C5.0,C5.0)、树扩展朴素贝叶斯算法(tree augmented Bayes network,TAN)和反向传播人工神经网络算法(back-propagation artificial neural network,BP-ANN)3种算法进行了评价。结果 C5.0、TAN和BP-ANN 3种算法构建的模型均适用于不同地域普洱茶的产地溯源。利用构建的模型经过外部样品预测集验证,以决策树算法C5.0构建的普洱茶产地溯源模型判别率达83.33%。结论化学计量学工具构建的普洱茶产地溯源模型将为普洱茶溯源工作提供技术保障,从而促进普洱茶产业的健康发展。
Objective To establish prediction model for identifying the different regions of Pu-erh tea origin information of Yunnan based on different elements of tea and chemical metrology software. Methods The content of 21 mineral elements and 15 rare earth elements in 107 tea samples from 2 origins was determined by inductively coupled plasma-mass spectrometry (ICP-MS). Chemical metrology software was used for data filtering, principal component analysis and origin identification model establishment. While the models for tea region authentication were established by Fisher linear discriminant analysis (FLD), decision tree 5.0 (C5.0), tree augmented Bayes network (TAN) and back-propagation artificial neural network (BP-ANN) algorithms and they were evaluated by comparing the discriminant accuracy. Results C5.0 model, TAN model and BP-ANN model were applied for geographical origin traceability of 2 Pu-erh tea regions. As applied by external validation samples, C5.0 model was suitable for geographical origin traceability of different regions, while the accuracy rate was 83.33%. Conclusion The origin traceability model of Pu-erh tea will provide a technical support for Pu-erh tea back work and promote the healthy development of Pu-erh tea industry.
出处
《食品安全质量检测学报》
CAS
2015年第9期3646-3653,共8页
Journal of Food Safety and Quality
基金
公益性(农业)行业专项(S212103046)
国家茶叶产品风险评估专项(GJFP2015005)~~
关键词
普洱茶
微量元素
原产地
模型构建
化学计量学工具
Pu-erh tea
trace elements
place of origin
model building
chemical metrology software
