zbMATH — the first resource for mathematics

Single sample modal identification of a nonstationary stochastic process. (English) Zbl 0557.93066
In this paper the authors show that it is possible to estimate the pole parameters of a Gauss-Markov process consistently from an increasing sample even when the white noise input has a time-varying covariance matrix. This, perhaps, surprising property follows from the observation that the pole parameters appear linearly in equations that can be solved by the realization algorithms. The paper is well written and the conclusions are amply documented.
Reviewer: J.Rissanen

93E12 Identification in stochastic control theory
62F12 Asymptotic properties of parametric estimators
62M10 Time series, auto-correlation, regression, etc. in statistics (GARCH)
93C99 Model systems in control theory
93C55 Discrete-time control/observation systems
93E10 Estimation and detection in stochastic control theory
93C57 Sampled-data control/observation systems
Full Text: DOI