Title :
Complexity measures for classes of neural networks with variable weight bounds
Author :
Finnoff, William
Author_Institution :
Siemens AG, Munich, Germany
Abstract :
The author derives complexity measures for classes of single-hidden-layer feedforward networks based on the capacity and metric entropy of a class of functions. Based on these measures, some deficiencies in commonly used complexity-penalty terms implemented to prevent overfitting are indicated
Keywords :
neural nets; capacity; complexity measures; complexity-penalty terms; metric entropy; neural networks; single-hidden-layer feedforward networks; variable weight bounds; Cost function; Current measurement; Entropy; Feedforward neural networks; Logistics; Neural networks; Pressing; Research and development; Stochastic resonance; Training data;
Conference_Titel :
Neural Networks, 1991. 1991 IEEE International Joint Conference on
Print_ISBN :
0-7803-0227-3
DOI :
10.1109/IJCNN.1991.170349
