Predicting Supercritical Extraction of St. John’s Wort by Simple Quadratic Polynomial Model and Adaptive Neuro-Fuzzy Inference System- Firefly Algorithm

Author(s): Ramin Tahmasebi Boldaji, Hossein Rajabi Kuyakhi

In this study, the applicability of the adaptive neuro-fuzzy inference system (ANFIS) and response surface methodology (RSM) was evaluated to forecasting the supercritical extraction of St. John’s Wort. In this case, the ANFIS model was optimized by the firefly algorithm (FFA) to develop the performance of the model. The accuracy of the models was investigated by comparing the result of the models with experimental data. The precision of the RSM model has been investigated using statistical analysis (P-value<0.05, F-value=148.13, R2=0,99 R2 Adjusted=0.98, R2 Predicted =0.96). The data transfer power was proved by the Box-Cox plot. Also, the polynomial model shows that it is much simpler and easier than other complex models. The results show, the ANFIS-FFA with R2=0.99, RMSE=1.79 and AARD%=2.79 have a high capability of predicting the St. John’s Wort extraction amount