Title :
Least squares identification with stopping
Author_Institution :
Dept. of Math., Wayne State Univ., Detroit, MI, USA
Abstract :
A stopping-time problem for least-squares identification is studied. The stopping rule is so determined that the recursive online algorithm is terminated if the iterates are inside an ellipsoidal confidence region with small volume, and with limited confidence coefficient close to one. A limit theorem is obtained for the stopped process by means of the weak convergence approach
Keywords :
convergence of numerical methods; least squares approximations; parameter estimation; ellipsoidal confidence region; least-squares identification; limit theorem; limited confidence coefficient; parameter identification; recursive online algorithm; stopping-time problem; weak convergence approach; Convergence; Ellipsoids; Least squares approximation; Least squares methods; Linear systems; Mathematics; Parameter estimation; Recursive estimation;
Conference_Titel :
Decision and Control, 1988., Proceedings of the 27th IEEE Conference on
Conference_Location :
Austin, TX
DOI :
10.1109/CDC.1988.194580
