二值三谱法评价复方甘草片质量一致性并与AI合成3D谱评价对比研究OA

Based on high performance liquid chromatography-diode array detector(HPLC-DAD)chromatographic data from 75 batches of compound liquorice tablets,this study proposed and validat-ed a 3D fingerprint constructed by AI-integrated fusion across 38 wavelengths in the 200-348 nm range.Combined with a binary triple fingerprint method to calculate Sm and Pm,this approach en-abled quality consistency evaluation.Simultaneously,nine pharmacologically active substances were quantitatively determined using standard curve method and ratio fingerprint method,with compara-tive analysis conducted.Results demonstrated that compared to the single-wavelength fingerprint at 220 nm,the 3D fingerprint maintained macro-qualitative consistency while significantly enhancing the discrimination power for Pm.Principal component analysis(PCA)and radar charts indicated its superior sensitivity in distinguishing batches and manufacturers.Furthermore,the ratio fingerprint method and standard curve method showed high consistency(r>0.99)in quantifying the nine target components,validating the reliability of the quantitative strategy.The 3D fingerprint integrates de-tection information from 38 wavelengths,enabling a comprehensive and sensitive characterization of the chemical composition and quality variations in compound liquorice tablets.This significantly en-hances the ability to distinguish between batches and manufacturers.Combined with the binary triple fingerprint method,this approach provides a big data-driven consistency monitoring method for herb-al medicine quality in traditional Chinese medicine consistency evaluation and standardized control.It offers empirical support and a scalable technical pathway for quality monitoring research based on artificial intelligence and large-scale chromatographic data.