Nonlinear system identification with composite relevance vector machines

Abstract:

Nonlinear system identification based on relevance vector machines (RVMs) has been traditionally addressed by stacking the input and/or output regressors and then performing standard RVM regression. This letter introduces a full family of composite kernels in order to integrate the input and output information in the mapping function efficiently and hence generalize the standard approach. An improved trade-off between accuracy and sparsity is obtained in several benchmark problems. Also, the RVM yields confidence intervals for the pbkp_redictions, and it is less sensitive to free parameter selection. © 2007 IEEE.

Año de publicación:

2007

Keywords:

• Relevance vector machine (RVM)
• Nonlinear system identification
• Composite kernels

Fuente:

scopus