The presented work discusses the development of a rapid and precise analytical protocol using near infrared spectroscopy combined with multivariate data analysis to authenticate, detect and quantify most of the commonly encountered plant adulterants used in fraud of saffron stigmas including safflower, pomegranate fruit peel, calendula flower, paprika, curcuma, hibiscus, saffron stamens and exhaustively-extracted saffron stigmas. A Soft Independent Modelling of Class Analogies (SIMCA) model was constructed for authentication of saffron stigmas with 100% sensitivity and a Partial Least Squares-Discriminant Analysis (PLS-DA) model was successfully utilized for correct discrimination of unadulterated and intentionally adulterated saffron samples as it showed 100% sensitivity and 99% specificity. Quantitation of the amount of each individual adulterant was achieved through construction of partial least squares regression (PLSR) models accompanied by variable importance to projection (VIP) method for variable selection which revealed that bands in the spectral ranges 6000-5800 cm⁻¹ followed by 4600–4200 cm⁻¹ and 5400-5000 cm⁻¹ were the most important for correct prediction with detection limits as low as 1%. The models performance was tested using internal and external validation sets indicating their reliability in providing a useful quality assessment tool for saffron in an attempt to prevent its fraud.

提出的工作讨论了使用近红外光谱结合多元数据分析来鉴定，检测和定量用于藏红花柱头造假的大多数常见植物掺假物的开发，该分析方案的快速，精确分析方法包括藏红花，石榴果皮，金盏花。 ，辣椒粉，姜黄，芙蓉，藏红花雄蕊和详尽提取的藏红花柱头。构建了用于类比的藏红花柱头鉴定的具有100％敏感性的软独立类比建模（SIMCA）模型，并且成功地使用了偏最小二乘判别分析（PLS-DA）模型正确鉴定了纯净和故意掺假的藏红花样品，作为它显示出100％的敏感性和99％的特异性。^-1接着4600-4200厘米^-1和5400-5000厘米^-1是用于与检测限低至1％正确预测的最重要的。使用内部和外部验证集对模型的性能进行了测试，表明它们在为藏红花提供有用的质量评估工具以防止其欺诈方面的可靠性。