论文部分内容阅读
研究以椴树、槐花、葵花、苕子、油菜、枣花6种蜜源蜂蜜样品为研究对象,利用中红外光谱检测器对样品进行光谱扫描,结合线性判别分析方法(Linear discrimination analysis,LDA)建立蜂蜜品种鉴别模型。结果表明,在经过光谱预处理及主成分分析后,用LDA方法建立的蜂蜜品种鉴别模型,对训练集和测试集样品的预测识别准确率分别为97.59%和96.30%。上述参与训练与测试的全部样品作为训练集建立的模型对完全未参与建模的槐花蜜、油菜蜜样品预测其正确判别率分别为80.00%和94.87%。表明该方法在快速、准确鉴别蜂蜜品种具有可行性、实用性。
In this study, six honey samples from linden, Sophora japonica, sunflower, razor clam, rape and date flower were used as the research object. The samples were scanned by mid-infrared spectroscopy. Linear discrimination analysis (LDA) Establishment of honey variety identification model. The results showed that the accuracy of predicting the training set and the test set was 97.59% and 96.30% respectively after the pretreatment of the spectrum and the principal component analysis (LDA) to establish the honey variety identification model. All the samples that participated in the training and testing as the training set established the right discriminant rates of 80.00% and 94.87% for the samples of Sophora japonica and rapeseed that did not participate in the modeling at all. It shows that this method is feasible and practical in identifying honey varieties quickly and accurately.