基于化学成分的大米国家地理标志产品模式识别与分类研究

Pattern Recognition and Classification of Geographical Indication Rice Based on Chemical Components

综合运用主成分分析(PCA)和Fisher准则两类线性判别方法对广东增城丝苗米国家地理标志产品中无机元素成分、有机组成(氨基酸、脂肪、淀粉、蛋白质等)进行模式识别与聚类研究,并采用Fisher线性判别方法建立地标产品与非地标产品间的分类模型。结果表明:Fisher线性判别用于大米地理标志产品的分类与识别简单快速,在分析数据相对复杂的情况下,可快速建立分类模型并实现地理标志产品与非地理标志样本间的鉴定与识别。

Principal component analysis （PCA） and Fisher linear discriminant analysis were both applied for the pattern recognition and classification of Zengcheng mew rice from Guangdong province as a geographical indication （GI） product. A classification model between GI and non-GI products was established based on inorganic elements and organic compounds including amino acids, fat, starch and protein by Fisher linear discriminant analysis. The results showed that Fisher linear discriminant analysis was a simple and fast method for the classification and identification of geographical indication rice and could allow rapid establishment of a classification model for identification and reorganization between GI and non-GI products based on relatively complicated data.