Integrated GC–MS and FT‐NIRS Characterization of Metabolic Differences Between Fresh and Dried Fruits of Five Genus <i>Amomum</i>

Amomum is a diverse genus whose metabolites are dominated by volatile compounds; however, differences in metabolism levels between fresh and dried fruits remain unclear. Our results indicated that metabolite differences between fresh and dried fruits were expressed in relative content rather than in species. Using a traditional machine learning model analysis, we observed that the characteristic metabolites distinguishing fresh and dried fruits differed across the two states. A total of 63 differential metabolites were identified and subsequently evaluated for their anti‐inflammatory activity. Through public database screening and threshold analysis, 27 core anti‐inflammatory targets were identified. These targets were involved in the inflammatory therapeutic mechanism of Amomum by coordinating pathways, such as chemical carcinogenesis–receptor activation, lipid and atherosclerosis, and pathways in cancer. Furthermore, molecular docking results showed that the binding energies of two core targets with seven key metabolites were less than −3.2, indicating stable interactions. Finally, nonvolatile compounds from fresh and dried fruits were characterized via Fourier transform near‐infrared spectroscopy. The established projections to least squares discriminant analysis (PLS‐DA) model exhibited a strong capacity to differentiate dried fruits, achieving 100% accuracy in both the training and test sets. The partial least squares regression model was also effective in predicting the content of anti‐inflammatory compounds in Amomum plants, with an optimal residual predictive deviation value of 2.86.