Drying temperature-dependent profile of bioactive compounds and prediction of antioxidant capacity of cashew apple pomace using coupled Gaussian Process Regression and Support Vector Regression (GPR–SVR) model

Abstract

Crude extracts from cashew apple pomace (CAP) dried at different temperatures were used in High-Pressure Liquid Chromatography to quantify total alkaloids content (TAC), total flavanoids content (TFC), total saponin content (TSC) and total phenolics content (TPC). Diphenyl-1-picrylhydrazyl (DPPH) was used to determine the antioxidant capacity (AOC) of CAP. Fourier-Transformed Infrared Spectroscopy-Attenuated Total Reflectance (FTIR-ATR) was used to identify the functional groups present in the pomace. TAC, TFC, TSC and TPC were used as inputs to model AOC using Gaussian Process Regression (GPR), and Support Vector Regression (SVR) and a coupled model was developed using the residuals of GPR and SVR. It was found that increasing drying temperature decreased TAC, TFC, TPC and AOC but TSC increased. Both GPR and SVR predicted AOC with high accuracy. Drying CAP at lower temperature preserved more bioactive compounds hence high AOC; FTIR-ATR showed that CAP has good hydration capacity and contains majorly inorganic phosphates, aliphatic hydrocarbons and primary alcohols. Model coupling enhanced AOC prediction.

Keywords

• Bioactive compounds quantification
• Cashew apple pomace
• Coupled GPR–SVR
• Drying
• Gaussian Process Regression modeling
• Predicting antioxidant capacity