Title :
Model selection, stochastic complexity and badness amplification
Author :
Gerencser, László ; Baikovicius, Jimmy
Author_Institution :
Dept. of Electr. Eng., McGill Univ., Montreal, Que., Canada
Abstract :
The authors present a type of predictive stochastic complexity which penalizes overparametrization more heavily than its traditional counterparts. It forms the basis for a type of model order selection method for ARMA (autoregressive moving average) processes, which performs exceptionally well, as shown by extensive simulation results
Keywords :
identification; statistical analysis; stochastic systems; ARMA processes; badness amplification; model order selection; overparametrization; statistical analysis; stochastic complexity; stochastic systems; Autoregressive processes; Complexity theory; Equations; Estimation theory; MIMO; Parameter estimation; Polynomials; Recursive estimation; Stochastic processes; Stochastic systems; Structural engineering;
Conference_Titel :
Decision and Control, 1991., Proceedings of the 30th IEEE Conference on
Conference_Location :
Brighton
Print_ISBN :
0-7803-0450-0
DOI :
10.1109/CDC.1991.261768
