摘要
采用减压蒸馏法提取镇江香醋香气,结合紫外-可见光谱技术对不同厂生产的镇江香醋进行快速鉴别。选取了195,235,250,275和310 nm波长处的吸光值作为光谱曲线的特征值,用主成分分析法对恒顺产的镇江香醋和其它6个厂产的进行聚类分析。主成分分析表明:前三个主成分对不同厂生产的镇江香醋具有较好的聚类作用,可以定性分析镇江香醋。把选取的5个特征值作为BP神经网络的输入,镇江香醋不同生产厂作为BP神经网络的输出,通过7个厂生产的共108个样本的训练和学习,建立了3层BP人工神经网络定量分析模型,对未知的14个样本进行鉴别,预测识别准确率达到100%。同时,感官分析了7个厂生产的样品,结果表明恒顺产镇江香醋与其它6个厂的有明显区别,这与紫外-可见光谱分析结果相符。
A method to discriminate different factories of Zhenjiang aromatic vinegar by means of vacuum distillation and UV-visible spectroscopy （190-800 nm） was developed. A relation has been established between the absorption spectra and the aroma of Zhenjiang aromatic vinegar. The dataset consists of a total of 122 samples of Zhenjiang aromatic vinegar; five absorptions at 195, 235, 250, 275 and 310 nm wavelength were chose, which were applied as characteristic values of spectra plot. First, the data was analyzed with principal component analysis （PCA）. It appeared to provide the reasonable clustering of different factories of Zhenjiang aromatic vinegar. Meanwhile PCA compressed hundreds of spectral data into a small quantity of principal components which described the body of the spectra; the five absorptions were applied as inputs to a back propagation artificial neural network with four hidden layer. One hundred eight samples from seven different factories were selected randomly, then they were used to build BP-ANN model. This model has been used to predict the varieties of 14 unknown samples; the recognition rate of 100% was achieved. The organoleptic results demonstrated that there was different between Hengshun Zhenjiang aromatic vinegar and other six factories samples, which matched well with UV-visible spectroscopy analysis results.
出处
《中国调味品》
CAS
北大核心
2009年第3期112-117,共6页
China Condiment
基金
江苏省镇江市农业重点攻关项目(NY2005401)
关键词
紫外-可见光谱
镇江香醋
减压蒸馏
主成分分析
BP神经网络
鉴别
UV-visible spectroscopy
Zhenjiang aromatic vinegar
vacuum distillation
principal component analysis
back propagation neural network
discrimination
