论文部分内容阅读
利用德国PEN3电子鼻系统快速检测四种食醋陈化期。通过电子鼻采集食醋挥发性成分的响应值,利用主成分分析(PCA)、线性判别分析(LDA)、Fisher线性判别分析(FDA)和多层感知器神经网络(MLPNN)分析进行模式识别,结果表明:LDA分析效果优于PCA分析;并且随着陈化时间的延长,食醋的气味成分变化有增快的趋势,这种气味的变化规律与酸度的变化规律相符。用Fisher线性判别和多层感知器神经网络建立食醋陈化时间的预测模型,发现Fisher线性判别对凤翔醋、陇县醋、金台醋和渭滨醋陈化期的正确检测率分别为100%、100%、98%和100%;多层感知器神经网络对凤翔醋、陇县醋、金台醋和渭滨醋陈化期的正确检测率分别为100%、100%、96.92%和100%。由于正确检测率的高低得出电子鼻结合Fisher线性判别对食醋陈化期的监测结果优于多层感知器神经网络。
Rapid detection of four kinds of vinegar aging by German PEN3 electronic nose system. The response values of the volatile components of vinegar were collected by electronic nose. Pattern recognition was performed by principal component analysis (PCA), linear discriminant analysis (LDA), Fisher linear discriminant analysis (FDA) and multilayer perceptron neural network (MLPNN) The results showed that LDA analysis was better than PCA analysis. With the aging time, the odor components of vinegar tended to increase rapidly. The change of odor was in accordance with the change of acidity. Fisher’s linear discriminant and multilayer perceptron neural network were used to establish the prediction model of vinegar aging time. The correct detection rate of Fisher linear discriminant on the aging of Fengxiang vinegar, Longxian vinegar, Jintai vinegar and Weicheng vinegar was 100 %, 100%, 98% and 100%, respectively. The correct detection rate of multi-layer perceptron neural network for Fengxiang vinegar, Longxian vinegar, Jintai vinegar and Weicheng vinegar were 100%, 100%, 96.92% and 100%. Due to the correct detection rate, the electronic nose combined with Fisher linear discriminant is superior to multi-layer perceptron neural network in monitoring the aging of vinegar.