Title :
Adaptive nonlinear filtering with the support vector method
Author :
Mattera, Davide ; Palmieri, Francesco ; Haykin, Simon
Author_Institution :
Dipt. di Ing. Elettron. e delle Telecomun., Univ. degli Studi di Napoli Federico II, Naples, Italy
Abstract :
The recently introduced Support Vector Method (SVM) is one of the most powerful methods for training a Radial Basis Function (RBF) filter in a batch mode. This paper proposes a modification of this method for on-line adaptation of the filter parameters on a block-by-block basis. The proposed method requires a limited number of computations and compares well with other adaptive RBF filters.
Keywords :
adaptive filters; nonlinear filters; support vector machines; adaptive nonlinear filtering; on-line adaptation; radial basis function filter; support vector method; Accuracy; Approximation algorithms; Approximation methods; Convergence; Support vector machines; Time series analysis; Training;
Conference_Titel :
Signal Processing Conference (EUSIPCO 1998), 9th European
Conference_Location :
Rhodes
Print_ISBN :
978-960-7620-06-4
