Atractylodis Macrocephalae Rhizoma (AMR) is a traditional Chinese medicine used for gastrointestinal diseases. With increased demand, there are more and more places of cultivation for AMR. However, the quality of AMR varies from place to place, and there is no good way to distinguish AMR from different origins at present. In this paper, we determined the content of eight chemical components including 60% ethanol extracts, essential oil, polysaccharides, atractylenolides, and atractylone, obtained the color parameters of AMR powder by colorimetry, and odor information was captured by the electronic nose, all of which were combined with machine learning to establish a rapid origin traceability method. The results of the principal component analysis of the chemical components revealed that Zhejiang AMR has a high comprehensive score and overall better quality. The Kruskal–Wallis test demonstrated that there are varying degrees of differences in chemical composition and color parameters across the different origin. However, the accuracy of the classification model is low (less than 80%), making it difficult to distinguish between different origins of AMR. The electronic nose demonstrated excellent classification performance in the traceability of AMR from different origins, with accuracy reaching more than 90% (PLS-DA: 96.88%, BPNN: 96.88%, PSO-SVM: 100%). Overall, this study clarified the quality differences of AMR among different origins, and a rapid and precise method combining machine learning was developed to trace the origin of AMR.

The rapid detection and quantification of microbial quantity and aflatoxin are crucial for food safety and quality. In order to achieve rapid detection, nutmeg with mildew, but with difficult-to-observe mildew characteristics, was selected as the research object. Its intrinsic component (dehydrodiisoeugenol) and exogenous noxious substances (the total number of colonies and aflatoxin B₁) were determined to clarify their changes during the mold process. Subsequently, electronic nose (E-nose) was employed to analyze the odor of nutmeg and was combined with six machine learning algorithms to establish a classification model for samples with different degrees of mold. Finally, three algorithms were chosen as the preferred options to establish the prediction models of indicator content, which can not only identify whether nutmeg is edible but also measure each index. The results demonstrate the enormous potential of E-nose for real-time detection for assessing food safety. In terms of qualitative analysis, the established classification model can achieve a more than 90% true positive rate, suggesting that E-nose could identify early mildew. In quantitative analysis, E-nose combined with Back Propagation Neural Network achieved the highest prediction accuracy, since the correlation coefficient between the predicted value and the measured value of aflatoxin B₁ is 0.9776, the TAMC is 0.9443, and the TYMC is 0.9685. This study provides a reference for the rapid and comprehensive quality evaluation of mildew-prone nutmeg, and it confirms that E-nose can be applied as a quick and simple technology.

Atractylodes macrocephala Rhizoma (AMR) is a frequently used medicinal herb for treating gastrointestinal disorders, with its quality influenced by factors such as origin and cultivation duration. Traditional quality control methods for AMR are time-consuming and invasive, making the development of faster and more efficient alternatives urgently needed. This study aims to utilize electronic nose (E-nose) and electronic tongue (E-tongue) to achieve the acquisition of odor–taste two-dimensional information of AMR. Integrating this approach with machine learning (ML) enables intelligent transformation from “experience-driven” to “data-driven” quality assessment, thereby developing a rapid and cost-effective quality control strategy for AMR. Feature-extraction and feature-selection techniques were employed to optimize back-propagation neural network (BPNN) classification and regression models for eight key quality markers, selecting the optimal feature subset. Additionally, nine machine-learning algorithms were applied with the optimal feature subset to establish classification models for different AMR grades and quantitative regression models for eight components based on E-nose and E-tongue data. The results demonstrated that the E-tongue combined with the k-nearest neighbors (KNN) algorithm could achieve a rapid classification of AMR grades with an accuracy of 95.56%. It also successfully predicted the contents of the extract, volatile oil, polysaccharides, atractylenolide I, atractylenolide II, atractylenolide III, bis-atractylenolide, and atractylone, with the test set’s coefficient of determination (R²) values of 0.8874, 0.8313, 0.9628, 0.8406, 0.8736, 0.8532, 0.7758, and 0.8101, respectively. In conclusion, this study provides a comprehensive and rapid solution for AMR grade classification and quality evaluation, significantly improving efficiency compared with traditional methods. This strategy holds substantial promise for real-world applications, as it enables a high-throughput, non-destructive screening of AMR in settings such as post-harvest processing and market quality surveillance, thereby supporting the sustainable and intelligent development of the herbal medicine industry.