Title :
Taxonomy of neural transfer functions
Author :
Duch, Wtodzistaw ; Jankowski, Norbert
Author_Institution :
Dept. of Comput. Methods, Nicholas Copernicus Univ., Torun, Poland
Abstract :
The choice of transfer functions may strongly influence complexity and performance of neural networks used in classification and approximation tasks. A taxonomy of activation and output functions is proposed, allowing to generate many transfer functions. Several less-known types of transfer functions and new combinations of activation/output functions are described. Functions parameterize to change from localized to delocalized type, functions with activation based on nonEuclidean distance measures, bicentral functions formed from pairs of sigmoids are discussed
Keywords :
approximation theory; computational complexity; neural nets; pattern classification; transfer functions; activation functions; activation/output function combinations; approximation; bicentral functions; neural network complexity; neural network performance; neural transfer function taxonomy; non-Euclidean distance measures; nonEuclidean distance measures; pattern classification; sigmoid pairs; Adaptive systems; Approximation methods; Multilayer perceptrons; Neural networks; Neurons; Statistical analysis; Taxonomy; Transfer functions;
Conference_Titel :
Neural Networks, 2000. IJCNN 2000, Proceedings of the IEEE-INNS-ENNS International Joint Conference on
Conference_Location :
Como
Print_ISBN :
0-7695-0619-4
DOI :
10.1109/IJCNN.2000.861353
