LAUSR

Abstract In this paper, three different data-driven algorithms were employed including two nonlinear models (Artificial neural network (ANN) and Adaptive neuro-fuzzy inference system (ANFIS)) and a classical linear model (Multilinear regression analysis (MLR)) for the simulation of response surface for methyclothiazide (M) and amiloride (A) considered as (K’or k) modeling in HPCL using pH and composition of mobile phase (methanol) as the corresponding input variables. The experimental and simulated results were evaluated based on five different performance efficiency criteria namely; determination coefficient (R2), root mean square error (RMSE), correlation coefficient (R), mean square error (MSE) and mean absolute percentage error (MAPE). The obtained results demonstrated the promising ability of ANN and ANFIS over MLR models with average R-values of 0.95 in both training and testing phases. The results also indicated that, with regard to the percentage error, ANN and ANFIS models outperformed the MLR model and increased the accuracy up to 6% and 8%, respectively for K’ ​ (M) simulation, while for K’ (A), ANFIS increased the accuracy up to 5% and 4% for MLR and ANN, respectively. The overall results proved the reliability of artificial intelligence models (ANN and ANFIS) for the simulation of response surface optimization method.