Modeling and optimization of turmerone concentrations from fermented turmeric waste essential oil via ANN and statistical techniques

Abstract

Turmeric ( Curcuma longa L.) pulp waste is an underutilized byproduct rich in essential oils components, particularly turmerone, which has high therapeutic and commercial value. Despite its potential, its recovery from such waste has not been extensively explored. With fermentation – solid-state fermentation (SSF) as a preliminary process to increase yield with lower cost, the present study focused on predicting and optimizing the turmerone concentration in essential oils extracted from turmeric waste through spontaneous fermentation via two modeling approaches: artificial neural networks (ANNs) and response surface methodology (RSM). A central composite rotatable design (CCRD) was employed to assess the influence of two key variables, namely fermentation time and hydrodistillation duration, on the response turmerone concentration. An evaluation of the models through the coefficient of determination (R 2 ), mean square error (MSE), root mean square error (RMSE), mean absolute error (MAE), mean absolute percentage error (MAPE), mean percentage error (MPE), and Pearson's χ2 revealed that the ANN was superior (R 2 = 0.9952, MSE= 0.14, RMSE= 0.40, MPE= 0.20 and χ2 = 0.02) to the RSM model (R 2 = 0.9894, MSE= 0.19, RMSE= 0.44, MPE= −0.09 and χ2 = 0.12) in terms of prediction accuracy and estimation reliability. The performance advantages of the RSM model (MAE = 0.24 and MAPE = 2.32) over the ANN (MAE = 0.39 and MAPE = 3.07) are not significant enough to outweigh the ANN's overall superiority. On the basis of ANN predictions, the optimal process conditions were determined to be 10 days of fermentation and 2 h of hydrodistillation, leading to a maximum predicted turmerone concentration of 16 millimolar (mM). Furthermore, the statistical analysis highlighted the distillation time as the most influential factor on the turmerone concentration. This study shows ANN modeling enhances turmeric waste valorization and highlights fermentation’s significant impact on turmerone recovery efficiency in extraction cycles. • A sustainable method for valorizing turmeric waste into high-value essential oil via spontaneous fermentation was proposed. • ANN and RSM models were applied to optimize the turmerone concentration in turmeric waste essential oil. • ANNs demonstrated superior predictive accuracy compared with traditional statistical methods. • This is the first report of ANN application to determine the optimal concentration and extraction conditions for turmeric pulp waste.