Artificial intelligence-based models for the qualitative and quantitative prediction of a phytochemical compound...

Area of Science:

• Phytochemistry and Analytical Chemistry
• Computational modeling applied to natural products analysis.

Background:

• Isoquercitrin is a flavonoid with significant biological activities, including anticancer and anti-inflammatory properties.
• Accurate qualitative and quantitative analysis of isoquercitrin is crucial for its application.
• High-performance liquid chromatography (HPLC) is a key technique for analyzing phytochemicals like isoquercitrin.

Purpose of the Study:

• To predict the retention time (tR) and peak area (PA) of isoquercitrin using HPLC.
• To compare the predictive performance of nonlinear models (ANN, ANFIS, SVM) and a linear model (MLR).
• To identify the most accurate model for the qualitative and quantitative determination of isoquercitrin.

Main Methods:

• Utilized artificial neural network (ANN), adaptive neuro-fuzzy inference system (ANFIS), support vector machine (SVM), and multilinear regression analysis (MLR) models.
• Input variables included standard concentration, mobile phase composition (MP-A, MP-B), and pH.
• Model performance was evaluated using RMSE, MSE, DC, and CC metrics.

Main Results:

• All four models demonstrated capability in predicting isoquercitrin's qualitative and quantitative properties.
• ANFIS-M3 showed the highest prediction accuracy for peak area (PA).
• ANN-M3 emerged as the best model for predicting retention time (tR).

Conclusions:

• Computational models, particularly ANN and ANFIS, are reliable tools for isoquercitrin analysis.
• These models offer accurate qualitative and quantitative determination of bioactive compounds.
• The study highlights the potential of advanced modeling techniques in phytochemical analysis.