Title :
Comparison of recurrent methods of parameter estimation in macromodeling problems
Author :
Stepashko, V.S. ; Yefimenko, S.M.
Author_Institution :
Int. Res. & Training Centre of Informat ion Technol. & Syst. of NAS, MES of Ukraine, Kiev, Ukraine
Abstract :
The paper suggests new modifications of Gauss and Gramm-Schmidt methods for recurrent parameter estimation of sequentially complicated structures of regression models when modeling from data. The results of run time comparison of these algorithms with a recurrent algorithm based on the known bordering method are presented.
Keywords :
Gaussian processes; parameter estimation; regression analysis; Gauss method; Gramm-Schmidt method; data modeling; known bordering method; macromodeling problems; recurrent parameter estimation; regression models; sequentially complicated structures; bordering method; modeling from data; recurrent parameter estimation;
Conference_Titel :
Modern Problems of Radio Engineering, Telecommunications and Computer Science, 2008 Proceedings of International Conference on
Conference_Location :
Lviv-Slavsko
Print_ISBN :
978-966-553-678-9
