Sustainable beetroot bioactives extraction with NADES-MAE: A spider algorithm-optimized hybrid SVR-XGBoost model

Read full text
Authors: Mostafa Khajeh, Mansour Ghaffari-Moghaddam, Jamshid Piri, Afsaneh Barkhordar, Halil Şenol
Cited by: 3
References: 33

Abstract

Beetroot (Beta vulgaris) is rich in biologically active compounds, particularly phenolics and flavonoids with antioxidant properties. However, their efficient and sustainable extraction remains a significant challenge. This study developed a green extraction process for bioactive compounds from beetroot using natural deep eutectic solvents (NADES) and microwave-assisted extraction (MAE). A hybrid machine learning strategy based on support vector regression (SVR) and extreme gradient boosting algorithm (XGBoost) was employed for the optimization of parameters and prediction of bioactive compound yields and antioxidant activity. Thirty systematic experiments were conducted to evaluate the effects of microwave power (175–550 W), extraction temperature (50–95°C), extraction time (20–70 min), and mass–solvent ratio (30–60 mg/mL) on total phenolic content (TPC), total flavonoid content (TFC), and antioxidant activity (DPPH%). Two weight optimization approaches were tested in order to improve the hybrid SVR–XGBoost model's predictive performance: standard cross-validation and a spider optimization algorithm. Spider optimization method surpassed cross-validation consistently, generating higher predictive accuracy in the forms of test R 2 values of 0.75, 0.77, and 0.68 for TPC, TFC, and DPPH%, respectively. Variable-specific weight optimization revealed distinct extraction mechanisms in which SVR dominated TPC prediction (weight = 0.95) but required balanced ensemble methods for TFC (SVR: 0.59, XGBoost: 0.41). Validated models successfully identified distinct optimal extraction conditions for each of the three responses (TPC, TFC, and DPPH%), demonstrating the need for output-specific optimization. This study advances green extraction technologies by integrating eco-friendly solvents with intelligent process optimization, providing a robust platform for the sustainable extraction of bioactive compounds with strong potential for industrial applications in the nutraceutical and functional food industries.

Fields of Study & Subjects

Metaheuristic Optimization Algorithms Research Advanced Multi-Objective Optimization Algorithms Polysaccharides and Plant Cell Walls

Publication details

Journal / source
Applied Food Research
Publisher
Elsevier BV
Volume
6
Issue
1
Pages
101940–101940
Publication date
29 March 2026
Language
EN