Title :
A simple method to estimate the size of feedforward neural networks
Author_Institution :
Coll. of Comput. Sci. & Technol., Beijing Univ. of Technol., China
Abstract :
The concept of monotonic value is presented to predetermine the optimal or near optimal number of hidden neurons needed in a feedforward neural network with one hidden layer. The approximation of an estimated neural network to y=sin(2πx) is well simulated and the monotonic value of Xor is proven to be 2.
Keywords :
feedforward neural nets; function approximation; learning (artificial intelligence); estimated neural network approximation; feedforward neural networks; hidden neurons; monotonic value; optimal number predetermination; size estimation; Computer networks; Computer science; Computer simulation; Educational institutions; Feedforward neural networks; Intelligent networks; Laboratories; Neural networks; Neurons;
Conference_Titel :
Systems, Man and Cybernetics, 2003. IEEE International Conference on
Print_ISBN :
0-7803-7952-7
DOI :
10.1109/ICSMC.2003.1244594
