主成分空间聚类模型在酱油鉴伪中的应用研究

Application Research of Principal Component Space Clustering Model in the Identification of Soy Sauce

通过GC-MS分析酿造酱油,非酿造酱油的挥发性成分,共鉴定出53种化学成分,筛选出20种特征性成分。分别计算上述样品中特征成分的峰面积占比,数据进行标准化处理后,通过主成分分析在构建的主成分空间上进行投影,建立酿造酱油和非酿造酱油挥发性成分的主成分空间聚类模型,通过计算和比较待测样品与两个聚类空间重心的直线距离可实现非酿造酱油的鉴别。

The 53 kinds of volatile components of fermented soy sauce, non-fermented soy sauce （NFSS） were analyzed by GC-MS. A total of 20 characteristic components were identified. The peak area ratio of 20 kinds of volatile components was calculated. After data normalized treatment, principal component projected in the principal component space. The standard clustering space of fermented soy sauce and NFSS were established by clustering analysis technology. The NFSS identification could be realized by comparing the straight line distance between sample sand two clustering space centers.