Optimization of Phenolic and Flavonoid Compounds Extraction Process from Echium italicum L. Using Machine Learning Models

Optimization of Phenolic and Flavonoid Compounds Extraction Process from Echium italicum L. Using Machine Learning Models

Reza Mirzaiyan Dehkordi1 Amir Mohammad Tahoni2 Mohammad Mehdi Farshisaftbaf3 Daniyal Rezaeei4 Mohammad Reza Ebrahimgol5

1) Department of Agriculture, University of Shahrekord, Iran,
2) Department of Agriculture, University of Shahrekord, Iran,
3) Department of Agriculture, University of Shahrekord, Iran,
4) Department of Agriculture, University of Shahrekord, Iran,
5) Department of Agriculture, University of Shahrekord, Iran,

IOI Code : ABEB-ACZFZZ
Publication : 4th International Conference on Recent Advances in Engineering, Innovation and Technology - Belgium(eitconf.com)
Abstract :
    Background and Objective: Phenolic and flavonoid compounds are the main bioactive constituents in medicinal plants. This study aimed to optimize the extraction process of phenolic and flavonoid compounds from Echium italicum L. using machine learning models. Methods: A dataset of 120 extraction experiments was generated using a Central Composite Design. Extraction time (15–120 min), temperature (25–80°C), ethanol concentration (20–80%, v/v), and solvent-to-solid ratio (10:1–50:1 mL/g) were used as input variables. Polynomial Regression and Random Forest models were developed to predict TPC and TFC. Model performance was assessed using the coefficient of determination (R^2), root mean square error (RMSE), and mean absolute error (MAE). Results: Total phenolic content ranged between 2.5 and 5.4 mg gallic acid equivalent per gram dry weight, and total flavonoid content ranged from 1.7 to 4.1 mg quercetin equivalent per gram dry weight. The Polynomial Regression model achieved R² = 0.88 for phenolic compounds and R² = 0.87 for flavonoids. The Random Forest model showed better performance with R² = 0.92 for phenolic compounds and R² = 0.91 for flavonoids. Conclusion: Machine learning models, particularly Random Forest, can effectively predict and optimize the extraction process of phenolic and flavonoid compounds from Echium italicum L. This approach can reduce experimental costs and time in phytochemical studies.

Keywords : echium italicum machine learning phenolic compounds flavonoids extraction optimization random forest polynomial regression