A new order estimation technique for time series modeling
Michael Harold Davis, Wei Xing Zheng · IEEE Transactions on Automatic Control · 1997
A new approach to estimating the order of the autoregressive moving average model is proposed, which is based on the approximate stochastic realization introduced in Davis and Fotopoulos (1991). The present approach is attractive because overparameterization-a very common problem in order determination-is avoided successfully. Simulation results are included to illustrate the effectiveness of the proposed order estimation approach.