Retention time modeling of short-chain aliphatic acids in aqueous ion-exclusion chromatography systems under several conditions using computational intelligence methods (artificial neural network and adaptive neuro-fuzzy inference system)

Abstract

In this paper, computational intelligence methods such as Artificial Neural Network (ANN) and Adaptive Neuro-Fuzzy Inference System (ANFIS) were used to model the effective parameters on retention behaviour of short-chain aliphatic carboxylic acids (C1-C6) and predict their retention time. For this purpose, seven factors, in three different categories, have been considered as input parameters in modeling and prediction of retention time: (i) operational parameters (temperature, H2SO4 normality as a mobile phase concentration, and flow rate); (ii) column characteristic properties (length of column, width of column, column’s particle size as the size of stationary phase); and (iii) acid type (number of carbon atoms in the analyte acid). Experimental data of available literatures were used as a data set to train and test Computational Intelligence (CI) methods. After developing the models, prediction ability of the models was checked with additional 18 data sets. The obtained results of the models showed that the proposed models are very precise, efficient and useful for fast and accurate prediction of the retention time for aliphatic carboxylic acids. Comparison of the two approaches indicated that both methods provide good results. Root mean square error of 0.0255 and 0.004, sum of square error of 0.148 and 0.001, and correlation factor of 0.978 and 0.998 for ANN and ANFIS, respectively, demonstrated that the proposed ANFIS model is more accurate in compared to the ANN model.