使用电子鼻传感器和化学计量学模型检测马哈鱼、黄花鱼、红鲷和弱鱼的分解

Detection of decomposition in mahi-mahi, croaker, red snapper, and weakfish using an electronic-nose sensor and chemometric modeling

使用电子鼻传感器和化学计量学模型检测马哈鱼、黄花鱼、红鲷和弱鱼的分解

Sanjeewa R. Karunathilaka¹， Zachary Ellsworth¹ ，Betsy Jean Yakes²

¹ Joint Institute for Food Safety and Applied Nutrition, University of Maryland, College Park, Maryland, USA
² Center for Food Safety and Applied Nutrition, U.S. Food and Drug Administration, College Park, Maryland, USA

Abstract:
This study evaluated an electronic-nose (e-nose) sensor in combination with support vector machine (SVM) modeling for predicting the decomposition state of four types of fish fillets:mahi-mahi, croaker, red snapper, and weakfish. The National Seafood Sensory Expert scored fillets were thawed, 10-g portions were weighed into glass jars which were then sealed, and the jars were held at approximately 30◦C to allow volatile components to be trapped and available for analysis. The measurement of the sample vial headspace was performed with an e-nose device consisting of nanocomposite, metal oxide semiconductor (MOS), electrochemical, and photoionization sensors. Classification models were then trained based on the sensory grade of each fillet, and the e-nose companion chemometric software identified that eight MOS were the most informative for determining a sensory pass from sensory fail sample.
For SVM, the cross-validation (CV) correct classification rates for mahi-mahi, croaker, red snapper, and weakfish were 100%, 100%, 97%, and 97%, respectively. When the SVM prediction performances of the eight MOS were evaluated using a calibration-independent test set of samples, correct classification rates of 93–100% were observed. Based on these results, the e-nose measurements coupled with SVM models were found to be potentially promising for predicting the spoilage of these four fish species.

本研究评估了电子鼻（e-nose）传感器与支持向量机（SVM）模型相结合，以预测四种类型鱼片的分解状态：马哈鱼、黄花鱼、红鲷和弱鱼。国家海鲜感官专家评分的鱼片解冻，将10克的鱼片称重放入玻璃罐中，然后密封，罐内温度约为30℃以允许挥发性成分被捕获并可用于分析。使用由纳米复合材料、金属氧化物半导体（MOS）、电化学和光电离传感器组成的电子鼻装置测量样品瓶顶空气体。然后，根据每个鱼片的感官等级对分类模型进行训练，电子鼻辅助化学计量学软件确定，8个MOS对于从感官失败样本中确定感官合格最有效。

对于SVM，马哈鱼、黄花鱼、红鲷和弱鱼的交叉验证（CV）正确分类率分别为100%、100%、97%和97%。当使用独立于校准的样本测试集评估八个MOS的SVM预测性能时，观察到93–100%的正确分类率。基于这些结果，电子鼻测量与SVM模型相结合被发现在预测这四种鱼类的腐败方面具有潜在的前景

Practical Application:
This report describes the application of an electronicnose sensor as a potential rapid and low-cost screeningmethod for fish spoilage.
It could provide regulators and stakeholders with a practical tool to rapidly and accurately assess fish decomposition.

本报告描述了电子传感器作为一种潜在的快速和低成本的鱼类腐败筛选方法的应用。
它可以为监管机构和利益相关者提供一个快速准确评估鱼类分解的实用工具。

KEYWORDS
electronic-nose, seafood, decomposition, screening

电子鼻，海鲜，分解，筛选

*本文内容非商业广告，仅供专业人士参考!