Modeling the adsorption of chlorobenzene on modified bentonite using an artificial neural network

Abstract In this paper, a radial basis function (RBF) feed-forward network has been used to predict the adsorption capacity for chlorobenzene (CB) onto modified Algerian bentonite (HDTMA-bentonite). The input parameters used for training of the RBF model include contact time, amount of adsorbent, pH, and the initial concentrations of chlorobenzene. The adsorption capacity for chlorobenzene onto modified bentonite is considered as an output of the neural network. An RBF model is used to predict the behavior of the adsorption process with the Levenberg–Marquardt algorithm (LMA). The model uses a linear transfer function (purelin) at the output layer and a tangent sigmoid transfer function (tansig) in the hidden layer with four neurons. The values of the determination coefficient (R2 ​= ​0.984) and the root mean square error (RMSE ​= ​0.015) showed good prediction results. Hence, the RBF model as a predictive tool has a great capacity to estimate the effect of operational parameters on the adsorption process.