Quality identification of Amomi fructus using E-nose, HS-GC-IMS, and intelligent data fusion methods

<i>Amomi fructus</i> (AF) has been used for both medicinal and food purposes for centuries. However, issues such as source mixing, substandard quality, and product adulteration often affect its efficacy. This study used E-nose (EN) and headspace-gas chromatography-ion mobility spectrometry (HS-GC-IMS) to determine and analyze the volatile organic compounds (VOCs) in AF and its counterfeit products. A total of 111 VOCs were detected by HS-GC-IMS, with 101 tentatively identified. Orthogonal Partial Least Squares-Discriminant Analysis (OPLS-DA) identified 47 VOCs as differential markers for distinguishing authentic AF from counterfeits (VIP value >1 and <i>P</i> < 0.05). Based on the E-nose sensor response value and the peak volumes of the 111 VOCs, the unguided Principal Component Analysis (PCA), guided Principal Component Analysis-Discriminant Analysis (PCA-DA), and Partial Least Squares-Discriminant Analysis (PLS-DA) models were established to differentiate AF by authenticity, origin, and provenance. The authenticity identification model achieved 100.00% accuracy after PCA analysis, while the origin identification model and the provenance identification model were 95.65% (HS-GC-IMS: PLS-DA) and 98.18% (HS-GC-IMS: PCA-DA/PLS-DA), respectively. Further data-level fusion of E-nose and HS-GC-IMS significantly improved the accuracy of the origin identification model to 97.96% (PLS-DA), outperforming single-source data modeling. In conclusion, the intelligent data fusion algorithm based on E-nose and HS-GC-IMS data effectively identifies the authenticity, origin, and provenance of AF, providing a rapid and accurate method for quality evaluation.