Title :
Rough Set Theory based Neural Network Architecture
Author :
Chandana, Sandeep ; Mayorga, Rene V.
Author_Institution :
Regina Univ., Regina
Abstract :
This paper proposes a novel method of combining rough concepts with neural computation. The proposed new rough neuron consists of, one lower bound neuron and another boundary neuron. The combination is designed in such a way that the boundary neuron deals only with the random and unpredictable part of the applied signal. Such architecture effectively prunes the search space for the respective constituent neurons based on the certain and uncertain behaviors. This division results in an improved rate of error convergence in the back propagation of the neural network along with an improved parameter approximation during the network learning process. Preliminary structures of the rough neural network along with some testing results have been presented. Further, performance comparisons with some of the prevalent designs have been done.
Keywords :
approximation theory; learning (artificial intelligence); neural nets; rough set theory; network learning process; neural network architecture; parameter approximation; rough neuron; rough set theory; uncertain behaviors; Biological neural networks; Computer architecture; Convergence; Intelligent systems; Neural networks; Neurons; Rough sets; Set theory; Signal design; Testing; Boundary Neuron; Output Excitation Factor; Rough Approximation; Rough Neural Computing;
Conference_Titel :
Neural Networks, 2006. IJCNN '06. International Joint Conference on
Conference_Location :
Vancouver, BC
Print_ISBN :
0-7803-9490-9
DOI :
10.1109/IJCNN.2006.246815