Title :
Locally regularised orthogonal least squares algorithm for the construction of sparse kernel regression models
Author_Institution :
Dept. of Electron. & Comput. Sci., Southampton Univ., UK
Abstract :
The paper proposes to combine orthogonal least squares (OLS) model selection with local regularisation for efficient sparse kernel data modelling. By assigning each orthogonal weight in the regression model with an individual regularisation parameter, the ability for the OLS model selection to produce a very parsimonious model with excellent generalisation performance is greatly enhanced.
Keywords :
Hessian matrices; data models; least squares approximations; statistical analysis; Hessian matrix; data modelling; learning procedure; local regularisation; orthogonal least squares algorithm; sparse kernel regression models; Bayesian methods; Computer science; Diversity reception; Iterative algorithms; Kernel; Learning systems; Least squares methods; Matrix decomposition; Optimization methods; Signal to noise ratio;
Conference_Titel :
Signal Processing, 2002 6th International Conference on
Print_ISBN :
0-7803-7488-6
DOI :
10.1109/ICOSP.2002.1180013