Comparison Of Information Criterion On Identification Of Discrete-Time Dynamic System

Information criterion is an important factor for model structure selection in system identification. It is used to determine the optimality of a particular model structure with the aim of selecting an adequate model. A good information criterion not only evaluate predictive accuracy but also the parsimony of model. There are many information criterions those are widely used such as Akaike Information Criterion (AIC) corrected Akaike Information Criterion (AICc) and Bayesian Information Criterion (BIC). Another information criterion suggesting use of logarithmic penalty, named as Parameter Magnitude-based Information Criterion (PMIC) was also introduced. This study presents a study on comparison between AIC, AICc, BIC and PMIC in selecting the correct model structure for simulated models. This shall be tested using computational software on a number of simulated systems in the form of discrete-time models of various lag orders and number of term/variables. As a conclusion, PMIC performed in optimum model structure selection better than AIC, AICc and BIC