逐步判别分析在烤烟产地鉴定中的应用

Application of Stepwise Discriminant Analysis in Identification of Tobacco Producing-area

以我国主产烟区烤烟主要化学成分为指标,采用逐步判别分析的方法对不同产地的烤烟样品进行判别分析,并建立判别函数。研究结果表明:总植物碱、总糖、淀粉、钾、氯和石油醚提取物6个主要化学成分进入判别函数,并建立判别模型。采用交互验证的方法对判别模型进行回判,对原样品判对率为79.7%,对预测样品判对率为86.7%。基于烤烟主要化学成分的判别分析模型有较高的稳定性,可以用于烤烟产地的鉴定。

The primary chemical compositions of flue - cured tobacco leaves from main tobacco - planting areas of China were used as indexes, the flue - cured tobacco samples from different producing - areas were analyzed by using the method of stepwise discriminant analysis, and the discriminant fimction was established. The results indicated that six chemical compositions such as total alkaloids, total sugar, starch, potassium, chlorine and petroleum ether extract were included in the discriminant function and the discriminant model was also established. The diseriminant model was re - discriminated by using the method of mutual test, its discrimination accuracy was 79.7% for the original samples, and was 86.7% for the unknown samples. The discriminant analysis model based on the primary chemical compositions of flue - cured tobacco had better stability, and it could be used to identify the producing - area of tobacco.