This study aimed to develop an artificial intelligence (AI)-based UV–Vis spectrophotometric method for simultaneously quantifying naringin and naringenin in Thymus canoviridis methanol extract. Conventional spectrophotometric techniques often face challenges in analyzing mixtures of compounds with overlapping spectra, making accurate quantification difficult. This limitation calls for advanced modelling approaches that incorporate machine learning to enhance prediction accuracy while aligning with green chemistry principles. Binary mixtures of naringin and naringenin were prepared at concentrations ranging from 5 to 40 μg/mL. UV–Vis spectra were collected over the 240–450 nm range. The trained machine learning models included Support Vector Regression (SVR), Linear Regression, Lasso Regression, Ridge Regression, and Elastic Net. These models were used to predict the concentrations of both flavonoids from their spectral data. Performance was evaluated using Mean Absolute Error (MAE) and the coefficient of determination (R2). The Ridge Regression model showed superior predictive performance, achieving MAEs of 0.882 for naringin and 0.378 for naringenin, with R2 values of 0.9912 and 0.9984, respectively. The model was successfully applied to the Thymus canoviridis extract, enabling precise quantification of both compounds without chemical separation. The method’s green chemistry profile was assessed using the AGREE and ComplexGAPI tools. AGREE scored it 0.72, reflecting strong green chemistry compliance, while ComplexGAPI revealed an overall environmentally favorable profile with some moderate environmental burdens related to solvent use. This study presents a green UV–Vis spectrophotometric method enhanced by AI for the simultaneous quantification of naringin and naringenin in plant extracts. By leveraging AI for spectral differentiation, the method achieves high accuracy and reliability without requiring chemical separation. This approach represents a significant advancement in plant metabolite analysis, facilitating high-throughput screening in phytochemical studies. The method's minimal solvent use, low waste generation, and energy efficiency make it a sustainable alternative to traditional techniques, suited for environmentally conscious phytochemical analysis.
MİLOĞLU et al. (Fri,) studied this question.