基于电子鼻和人工神经网络对鱼露产地的区分与识别

Distinction and Recognition of the Fish Sauce Origin Based on Electronic Nose and Artificial Neural Network

为建立一种快速无损检测鱼露产地的方法,采购不同产地(泰国、广东、越南、实验室自制)的鱼露,利用电子鼻检测其风味物质。通过主成分分析(PCA)对不同产地的鱼露进行区分,人工神经网络(ANN)对鱼露的产地进行预测。结果显示:在主成分析图上,总贡献率为98.37%,不同产地鱼露能够被较好地区分;负荷加载分析(Loading)表明P30/1(对烃类、氨类和乙醇类物质敏感)和LY2/GH(对氨类和胺类化合物敏感)传感器在区分鱼露产地中起主要作用,径向基人工神经网络(RB)和梯度下降人工神经网络(GD)对鱼露产地进行预测的准确率分别为95.56%和91.04%,表明采用电子鼻技术鉴别不同产地的鱼露是可行的。

Establish a rapid and non-destructive detection method for fish sauce production, purchasing fish sauce from different areas（Thailand, Guangdong, Vietnam, laboratory self-made）, and using electronic nose to detect the flavor. Use principal component analysis（PCA） to distinguish different origins of fish sauce, then using artificial neural network（ANN） to predict the origins of fish sauce. The results show that in PCA plot, the total contribution rate is 98. 37%, different origins of fish sauce could be separated well; loading analysis results show that the sensor P30/1 （hydrocarbons, ammoaia and ethanol） and LY2/GH（ammonia and amines） play a main role in distinguishing fish sauce production, and the predicted values of origins of fish sauce in both radial basis（RB） and gradient descent（GD） are 95.56% and 91.04%. Therefore, using the electronic nose technology to identify different origins of fish sauce is feasible.