Guaranteed estimation of linear regression parameters under dependent disturbances

Stationary stochastic processes (60G10) Linear regression; mixed models (62J05) Estimation and detection in stochastic control theory (93E10) Identification in stochastic control theory (93E12) Least squares and related methods for stochastic control systems (93E24)

Review text

The authors consider a regression model of the type \[ x_n= \sum^m_{j=1} \theta_j\varphi_j(n)+ \xi(n),\quad n\geq 1, \] where \(\varphi_i(n)\) are known deterministic functions, \(\xi(n)\) is the disturbance sequence, \(\theta_i\) are unknown parameters (deterministic trend). Mean-square guaranteed estimators of the regression parameters are constructed under the assumption that the disturbance is described by an autoregression process with unknown parameters and distribution. The method of sequential least-squares estimations of the parameters of the autoregression process is used for the solution to this problem. The authors construct the procedure of simultaneous estimation of the autoregression parameters and the deterministic trend, with given mean-square accuracy. Periodic trend is considered as a particular case. A limiting relation between the number of observations in the sequential procedure and estimation accuracy is derived.

Recommendations

Cited in

(1)

On guaranteed parameter estimation of a multiparameter linear regression process

This page was built for publication: Guaranteed estimation of linear regression parameters under dependent disturbances

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1286550)