Title :
Support vector regression and NARMAX system identification
Author :
Drezet, P. ; Harrison, R.F.
Author_Institution :
Dept. of Autom. Control & Syst. Eng., Univ. of Sheffield, Sheffield, UK
fDate :
Aug. 31 1999-Sept. 3 1999
Abstract :
Support Vector Regression (SVR) is a flexible regression method, which can be applied directly to NARMAX system identification models. SVR is a one-step convex optimisation process which attempts to maximise generalisation performance. This paper compares SVR performance with that of multi-layer perceptrons and radial basis function networks for varying numbers of time lags included in the model.
Keywords :
convex programming; identification; nonlinear control systems; regression analysis; support vector machines; NARMAX system identification; SVR; convex optimisation process; support vector regression; Computational modeling; Data models; Mathematical model; Neural networks; Optimization; Support vector machines; Training; NARMAX; Nonlinear Regression; Support Vector Regression; System Identification;
Conference_Titel :
Control Conference (ECC), 1999 European
Conference_Location :
Karlsruhe
Print_ISBN :
978-3-9524173-5-5
