Title :
Adaptive transfer functions [backpropagation neural nets]
Author :
Goulding, John R.
Author_Institution :
Portland State Univ., OR, USA
Abstract :
The author details the approach and methodology used to build adaptive transfer functions in a feedforward backpropagation neural network and provides insight into the structure-dependent properties of using nonscaled analog inputs. The results of using adaptive transfer functions are shown to outperform conventional architectures in the implementation of a mechanical power transmission gearbox design expert system knowledge base
Keywords :
artificial intelligence; expert systems; learning systems; neural nets; adaptive transfer functions; feedforward backpropagation neural network; mechanical power transmission gearbox design expert system knowledge base; nonscaled analog inputs; structure-dependent properties; Backpropagation; Decoding; Encoding; Expert systems; Gears; Manufacturing; Mechanical power transmission; Neural networks; Process design; Transfer functions;
Conference_Titel :
Neural Networks, 1991., IJCNN-91-Seattle International Joint Conference on
Conference_Location :
Seattle, WA
Print_ISBN :
0-7803-0164-1
DOI :
10.1109/IJCNN.1991.155396
