基于电子鼻技术的传统手工牦牛酥油品质评价

The evaluation on the quality of traditional yak ghee by electronic nose

【目的】评价传统手工牦牛酥油的品质,为食品工业的检验和产品品质优化提供参考依据。【方法】以16家牧户的传统手工牦牛酥油为原材料,采用最小显著性差异法(least significant difference,LSD)研究不同牧户酥油样品酸价、过氧化值和气味信息间的差异,采用主成分分析(principal component analysis,PCA)、典则判别分析(canonical discriminant analysis,CDA)和多层感知模型(multi-layer perceptron,MLP)定性判别及定量分析酥油产地和来源。【结果】16份牦牛酥油酸价(4.35~4.71 mg/g)均高于国标要求,过氧化值(0.0000~0.0075 g/100g)均在要求范围内;不同牧户牦牛酥油品质有差异。PCA分析可把16户牦牛酥油样品来源进行初步区分,CDA分析结果中各样品的数据点更聚集、样品组间距离变大,区分效果强于PCA,但仍有较多数据点重叠。采用MLP可定量判别传统手工牦牛酥油产地并追溯其来源,MLP 10-1-2网格结构结果显示,碌曲和玛曲两个产地的牦牛酥油类别效果平均值:准确率、精确率、灵敏度、特异度以及曲线下方面积分别为99.51%、99.53%、99.53%、99.50%和99.52%;MLP 10-1-16网格结构显示,16家牧户的牦牛酥油类别效果平均值:准确率、精确率、灵敏度、特异度以及曲线下方面积分别为98.97%、92.18%、92.23%、99.45%和95.84%。【结论】基于电子鼻技术可评价传统手工牦牛酥油品质。

【Objective】The objective of this study was to evaluate the quality of traditional yak ghee and provide reference for food industry inspection and product quality optimization.【Method】The acid value,peroxide value,and odor information of yak ghee samples were analyzed using the Least Significant Difference(LSD)test.The place of production and origin of yak ghee were qualitatively and quantitatively analyzed using Principal Component Analysis(PCA),Canonical Discriminant Analysis(CDA)and Multi-Layer Perceptron(MLP).【Result】The acid value(4.35~4.71 mg/g)was found to be higher than the national standard,while the peroxide value(0.000~0.0075 g/100g)fell within the required range.The quality of yak ghee varied among different families.PCA analysis successfully distinguished the sources of yak ghee samples from 16 families.In CDA analysis,the data points of each sample were more concentrated,resulting in larger distances between sample groups and a stronger discrimination effect compared to PCA.However,there were still some overlapping data points.MLP was employed for quantitative identification of the origin of traditional handmade yak ghee and traceability.The MLP 10-1-2 grid structure showed that the average accuracy,accuracy,sensitivity,specificity,and area under the curve for yak ghee in Luqu and Maqu were 99.51%,99.53%,99.53%,99.50%,and 99.52%,respectively.The MLP 10-1-16 grid structure showed that the average accuracy,accuracy,sensitivity,specificity,and area under the curve for yak ghee from 16 herdsmen were 98.97%,92.18%,92.23%,99.45%,and 95.84%,respectively.【Conclusion】Based on the findings of this study,the quality of traditional yak ghee can be effectively evaluated using E-nose technology.