System identification for block-ice process: Comparison between linear and neural network models

Abstract

This paper presents an application of system identification techniques for block-ice process. In particular, we compare two parametric models, namely, linear and neural network models. Linear models are built with Auto-Regressive model with exogenous inputs (ARX structure). Feedforward neural networks (NNARX structure) are constructed using the information provided by the linear ARX models. In addition, the Optimal Brain Surgeon (OBS) method is employed to prune the neural network models. Comparing the results obtained from the ARX and NNARX models, it is clearly seen that NNARX models yield better performance than ARX models.