An Optimized Detection Method for Chinese Red Huajiao Geographical Origin Determination, Based on Electronic Tongue and Ensemble Recognition Algorithm
An Optimized Detection Method for Chinese Red Huajiao Geographical Origin Determination, Based on Electronic Tongue and Ensemble Recognition Algorithm
复制标题
DOI：
10.1007/s12161-022-02285-4
复制
发表时间：
2022-04
影响因子：
2.9
通讯作者：
Yuezhong Mao;Shiwen Cheng;B. Shi;Lei Zhao;Shiyi Tian;Houyin Wang
中科院分区：
文献类型：
--
作者：
Yuezhong Mao;Shiwen Cheng;B. Shi;Lei Zhao;Shiyi Tian;Houyin Wang
红花椒是中国最有价值、用途最广的花椒品种，其品质因地域而异。本研究的目的是利用电子舌和集成识别算法建立一种优化的中国红花角原产地检测方法。利用电子舌对6种不同来源的样品进行检测，分别采用PCA-DI法和滤纸包裹法得到最优特征值和过滤后的特征值。采用5类集成识别算法中的22种模型分析方法进行地理来源识别。与传统的PCA-DI方法相比，滤波器法的总识别率最高，提高了6.1%。此外，集成支持向量机的识别能力最好，平均准确率为96.6%。因此，将电子舌头、滤纸和集成支持向量机相结合的优化方法将有可能应用于中国红花椒的产地和质量判定。
Red huajiao was the most valuable and widely used ofZanthoxylumspecies in China, and its quality was differed from the geographical region. The aim of this study was to establish an optimized detection method for Chinese red huajiao origin determination by using the electronic tongue and ensemble recognition algorithm. Six different origins of samples were detected by the electronic tongue; the optimized characteristic value and filtered characteristic value were obtained by the PCA-DI method and Filter-Wrapper method. Twenty-two kinds of model analysis methods that belonged to five categories of ensemble recognition algorithms were used to recognize the geographical origin. Compared to the traditional PCA-DI method, the best total recognition accuracy rate of Filter-Wrapper method was better with an increase of 6.1%. Furthermore, the ensemble SVM algorithm exhibited the best distinguishing ability with the average accuracy rate of 96.6%. Therefore, the optimized method that combined with electronic tongue, Filter-Wrapper, and ensemble SVM shall be promising for the Chinese red huajiao geographical origin and quality determination application.