Title :
A hybrid neural network/rule-based architecture for analogue function approximation
Author :
Curtis, K.M. ; Burniston, J.D.
Author_Institution :
Dept. of Math. & Comput. Sci., Univ. of the West Indies, Kingston, Jamaica
Abstract :
Investigations have been carried out into combining a rule-based system and an artificial neural network (ANN) to achieve a new computing structure for function approximation. Results are presented for the performance of the hybrid structure when applied to modelling a continuous nonlinear function, and are compared to the results obtained when modelling the function using only an ANN.
Keywords :
artificial intelligence; function approximation; knowledge based systems; neural net architecture; nonlinear functions; analogue function approximation; artificial neural network; nonlinear function; rule-based architecture; Artificial neural networks; Computer architecture; Computer networks; Function approximation; Hardware; MOSFET circuits; Neural networks; SPICE; Vectors; Voltage;
Conference_Titel :
SoutheastCon, 2003. Proceedings. IEEE
Print_ISBN :
0-7803-7856-3
DOI :
10.1109/SECON.2003.1268441
