EUROPEAN FOOD RESEARCH AND TECHNOLOGY, vol.248, no.12, pp.2995-3002, 2022 (SCI-Expanded)
In this study, colorimetric parameters of the red color emerging as a result of the novel modified Seliwanoff test were used for the quantitative determination of HMF in some syrup samples. Three different evaluation method (color space-based method, absorbance method and artificial neural network (ANN)) were used to determine HMF amounts of samples. While "da*" and absorbance value of red color were used to obtain a linear relationship with HMF concentration, L*, a*, b*, dL*, da*, db*, dE* were used as input variables in artificial neural network evaluation. Also HPLC method was carried out as a reference method. In addition to the absorbance value, 'da*' value also showed a linear relationship with the HMF concentration (da* = 0.0425 + 0.0612 x HMF concentration, ppm); correlation coefficient: 0.9980; p value of lack of fit: 0.579). When HMF amounts obtained by each method were compared with that by reference method HPLC, a strong correlation was found (color space-based method-HPLC, r: 0.996; absorbance method-HPLC, r: 0.997 and ANN-HPLC; r: 0.997). Among the methods, ANN evaluation method was more prominent in detecting low HMF concentrations (< 4 mg HMF/kg sample).