INVESTIGADORES
COUSSEAU Juan Edmundo
congresos y reuniones científicas
Título:
A PIECEWISE LINEAR DYNAMICAL FUNCTIONAL ARTIFICIAL NEURAL NETWORK (PWL-DFANN) FOR NONLINEAR ADAPTIVE TIME SERIES PREDICTION
Autor/es:
J.L. FIGUEROA; J. E. COUSSEAU; R.J.P DE FIGUEIREDO
Lugar:
Phoenix, AZ
Reunión:
Congreso; International Symposium on Circuits and Systems; 2002
Institución organizadora:
IEEE Circuits and Systems Society
Resumen:
A nonlinear adaptive time series predictor has been developedusing a PWL-DFANN network (Piecewise LinearDynamical Functional Artificial Neural Network)for its underlying model structure. This network isso called because it is a D-FANN the activation functionsof which are piecewise linear. The network hasbeen successfully used to model and predict an importantclass of highly dynamic and nonstationary signals,namely speech signals.
