A dual-fingerprint approach based on parameter systems for comprehensive quality evaluation of Belamcandae Rhizoma

In recent years, the widespread use of herbal medicines has made quality evaluation a critical issue. This study proposes three distinct methods to assess the similarity of dual fingerprints from Belamcandae Rhizoma for comprehensive quality consistency evaluation. For HPLC fingerprint similarity assessment, the super-information characteristic parameter system and modified Euclidean distance method were employed. The super-information characteristic parameters enabled the detection of subtle variations in HPLC chromatographic fingerprints, and their integration with principal component analysis provided an effective approach for discriminating quality similarity among samples. The modified Euclidean distance addressed several limitations inherent in the original Euclidean distance, enabling significantly enhanced performance with improved sensitivity for fingerprint similarity assessment. By integrating peak-specific weights with the modified Euclidean distance algorithm, a comprehensive evaluation of the HPLC fingerprints was achieved. K-means clustering combined with principal component analysis was applied to analyze the electrochemical oscillation fingerprints, providing the evaluation of sample quality similarity based on their overall chemical composition profiles. The three evaluation methods collectively demonstrated low similarity among the first nine batches and high consistency in the latter eight batches. Comparative analysis demonstrated that all three alternative methods exhibited superior sample discriminative capacity relative to the conventional cosine similarity method. This study established a comprehensive and reliable analytical strategy for quality consistency evaluation of Belamcandae Rhizoma, offering significant reference value for the quality control of the other herbal medicines.