Detection of adulteration in cherry tomato juices based on electronic nose and tongue: Comparison of different data fusion approaches
Detection of adulteration in cherry tomato juices based on electronic nose and tongue: Comparison of different data fusion approaches
复制标题
DOI：
10.1016/j.jfoodeng.2013.11.008
复制
发表时间：
2014-04-01
影响因子：
5.5
通讯作者：
Wang, Jun
中科院分区：
文献类型：
--
作者：
Hong, Xuezhen;Wang, Jun
采用七种方法对掺有不同程度的过熟番茄汁（0-30%）的新鲜樱桃番茄汁进行认证。考虑了两次电子鼻测量，结果表明在电子鼻测量之前使用干燥剂的预处理是不必要的。应用主成分分析（PCA）、因子F和逐步选择来构建融合数据集的特征。掺假水平的定性识别主要通过典型判别分析（CDA）和库支持向量机（Lib-SVM）进行。使用主成分回归 (PCR) 进行 pH 值和可溶性固形物含量 (SSC) 的定量校准。所有方法都表现出良好的分类性能，并且基于融合方法的预测性能优于单独使用电子鼻或电子舌的预测性能；然而，基于不同融合方法的分类和预测性能有所不同。这项研究表明，当使用适当的数据融合方法时，同时使用这两种仪器将比单独使用电子鼻或电子舌保证更好的性能。 (C) 2013 Elsevier Ltd. 保留所有权利。
Seven approaches were employed for authentication of fresh cherry tomato juices adulterated with different levels of overripe tomato juices: 0-30%. Two e-nose measurements were considered, and the result indicates that a pretreatment of using desiccant prior to e-nose measurement is unnecessary. Principle Component Analysis (PCA), factor F and stepwise selection were applied for feature construction of fusion datasets. Qualitative recognition of adulteration levels was mainly performed by Canonical Discriminant Analysis (CDA) and Library Support Vector Machines (Lib-SVM). Quantitative calibration with respect to pH and soluble solids content (SSC) was performed using Principle Components Regression (PCR). All the approaches presented well classification performances, and prediction performances based on fusion approaches are better than based on sole usage of e-nose or e-tongue; yet classification and prediction performances based on different fusion approaches vary. This study indicates that simultaneous utilization of both instruments would guarantee a better performance than individually utilization of e-nose or e-tongue when proper data fusion approaches are used. (C) 2013 Elsevier Ltd. All rights reserved.