基于气味变化规律采用电子鼻构建黄精霉变的快速判别模型

Establishment of a fast discriminant model with electronic nose for Polygonati Rhizoma mildew based on odor variation

通过对不同霉变程度的黄精进行气味指纹分析,探究黄精霉变过程中的气味变化规律与其霉变程度的关系,根据电子鼻的响应强度建立判别黄精是否霉变的快速判别模型。运用α-FOX3000电子鼻对不同霉变程度的黄精进行气味指纹分析,利用雷达图解析电子鼻挥发性化合物的主要贡献成分,通过偏最小二乘法判别分析(PLS-DA)、最邻近分类法(KNN)、序列最小最优化算法(SMO)、随机森林(random forest)、朴素贝叶斯(naive Bayes)对特征数据进行处理。根据电子鼻的雷达图可以得出,随着黄精样品的霉变程度不断加深,传感器T70/2、T30/1、P10/2的响应值一直在增强,说明霉变后的黄精产生了烷烃类和芳香族类化合物;根据PLS-DA,3类黄精样品可以很好地被区分,对传感器进行变量重要性分析,筛选出对分类贡献较大的传感器有5根:T70/2、T30/1、PA/2、P10/1、P40/1;在KNN、SMO、random forest、native Bayes 4个分类模型中,4个模型的分类准确率均达到90%以上,其中KNN为最佳分类模型,准确率为97.2%。霉变前后的黄精具有明显气味变化,产生了不同的挥发性有机物,这种差异可以被电子鼻识别,并建立了快速判别模型,为后续黄精霉变过程中挥发性物质种类和含量的研究提供方法参考。

The odor fingerprint of Pollygonati Rhizoma samples with different mildewing degrees was analyzed and the relationship between the odor variation and the mildewing degree was explored.A fast discriminant model was established according to the response intensity of electronic nose.Theα-FOX3000 electronic nose was applied to analyze the odor fingerprint of Pollygonati Rhizoma samples with different mildewing degrees and the radar map was used to analy ze the main contributors among the volatile organic compounds.The feature data were processed and analyzed by partial least squares discriminant analysis(PLS-DA),K-Nearest neighbor(KNN),Sequential Minimal Optimization(SMO),Random Forest(RF)and Naive Bayes(NB),respectively.According to the radar map of the electronic nose,the response values of three sensors,namely T70/2,T30/1,and P10/2,increased with the mildewing,indicating that the Pollygonati Rhizoma produced alkanes and aromatic compounds after the mildewing.According to PLS-DA model,Pollygonati Rhizoma samples of three mildewing degrees could be well distinguished in three areas.Afterwards,the variable importance analysis of the sensors was carried o ut and then five sensors that contributed a lot to the classification were screened out: T70/2, T30/1, PA/2, P10/1 and P40/1. Theclassification accuracy of all the four models (KNN, SMO, RF, and NB) was above 90%, and KNN was most accurate(accuracy: 97.2%). Different volatile organic compounds were produced after the mildewing of Pollygonati Rhizoma, and theycould be detected by electronic nose, which laid a foundation for the establishment of a rapid discrimination model formildewed Pollygonati Rhizoma. This paper shed lights on fur ther research on change pattern and quick detection of volatileorganic compounds in moldy Chinese herbal medicines.