Development of an SVR model for microwave-assisted aqueous two-phase extraction of isoflavonoids from Radix Puerariae

Abstract

In this study, simultaneous extraction and purification of isoflavonoids from Radix Puerariae was carried out by the microwave-assisted aqueous two-phase extraction technique (MAATPE). An ethanol/K₂HPO₄ system was used to form an aqueous two-phase system. The critical factors, namely, the concentrations of K₂HPO₄ and ethanol, solid–liquid ratio, extraction temperature, extraction time, and microwave power were investigated. Subsequently, the least squares support vector regression (LS-SVR) as a machine learning method was used to establish the model and predict extraction yield. The final conditions were concluded as follows: ethanol concentration 29.5% (w/w) and K₂HPO₄ concentration 24.9% (w/w), extraction temperature 79 °C, extraction time 8 min, and microwave irradiation power 600 W. Under the optimal condition, the extraction yield of total isoflavonoids reached 13.39%, which was similar to the predicted value of 13.48%. The recovery of isoflavonoids was 94.6%, and the yield of puerarin was 49.5 mg/g.

Keywords:

• Isoflavonoids
• microwave-assisted aqueous two-phase extraction
• Radix Puerariae
• support vector regression

Additional information

Funding

This work was supported by the National Natural Science Foundation of China (grant number 21006075); and the Natural Science Foundation of Hubei Province, China (grant number 2014CFA115).