基于多元素和脂肪酸指纹特征的中国北方大豆产地鉴别研究

Identification of soybean producing area in North China based on multi-element and fatty acid fingerprint characteristics

为探讨数据融合技术用于邻近区域大豆产地鉴别的可行性,在黑龙江省农垦九三管理局和绥化地区采集216份大豆样品,测定镁(Mg)、铝(Al)、磷(P)等13种矿物质元素含量和棕榈酸、硬脂酸、油酸等5种脂肪酸含量。分别利用矿物质元素、脂肪酸、数据级融合和特征级融合数据建立4种核函数的支持向量机(Support Vector Machine,SVM)模型。通过网格搜索算法结合五折交叉验证进行参数优化后,模型识别准确率分别提升至90.77%、92.31%、89.23%、95.38%。结果表明,特征级数据融合技术对邻近区域大豆产地鉴别效果显著,优于其他三种数据识别技术。采用特征级数据融合技术建立支持向量机产地鉴别模型,能够对邻近区域大豆产地进行准确、有效的区分,为地理标志产品保护技术提供了新的研究方向。

In order to explore the feasibility of data fusion technology in the origin identification of neighboring soybean producing area,216 soybean samples were collected from Heilongjiang Agricultural Reclamation Jiusan Administration and Suihua area,and the contents of 13 mineral elements such as magnesium(Mg),aluminum(Al)and phosphorus(P)and 5 fatty acids such as palmitic acid,stearic acid and oleic acid were determined.Support vector machine(SVM)with 4 kernel functions were established by mineral elements,fatty acids,data-level fusion and feature-level fusion data.After parameter optimization by grid search algorithm combined with 50% cross-validation,the recognition accuracy of the model increased to 90.77%,92.31%,89.23%and 95.38% respectively.The results showed that the feature-level data fusion technology had a significant effect on the origin identification of neighboring soybean producing area,which was superior to the other three data identification technologies.It could accurately and effectively distinguish soybean origin in neighboring areas by feature-level data fusion technology to establish support vector machine origin identification model,which provided a new research direction for geographical indication product protection technology.