机构地区: 中国农业大学理学院
出 处: 《分析化学》 2003年第7期843-845,共3页
摘 要: 采用傅里叶变换近红外光谱法测定大麦中蛋白质、淀粉、赖氨酸的含量 ,并用光谱影响值法(leverage)对异常值进行判断和处理。蛋白质、淀粉和赖氨酸含量近红外光谱分析模型的测定系数R2 分别为0 .985、0 .973和 0 .978;检验集的化学值与模型预测值的相关系数r分别为 0 .985 3、0 .96 4 4和 0 .9172 ,分析模型的预测相对标准偏差RSD分别为 4 .0 %、2 .4 %和 5 .4 % ,该结果可替代经典分析方法 ,满足农产品快速分析的需要。 The contents of protein, starch and lysine in barley were determined by Fourier transform near-infrared (FT-NIR) spectroscopy. Partial least square method was used to build a NIR model. Determination coefficient R-2 of protein, starch and lysine is respectively 0.985, 0.973, 0.978. The correlation coefficient r of independent validation set is 0.9853, 0.9644 and 0.9172 respectively, and relative standard deviation (RSD) is 4. 0 %, 2.4% 5.4% respectively. The outlier test of sample spectra and property value is also discussed in here.
关 键 词: 傅里叶变换近红外光谱法 测定 大麦 蛋白质 淀粉 赖氨酸 含量 偏最小二乘法 异常值
领 域: [农业科学]