Paper Info
Title
Feedforward sigmoidal networks--equicontinuity and fault-tolerance properties.
Abstract
Sigmoidal feedforward artificial neural networks (FFANNs) have been established to be universal approximators of continuous functions. The universal approximation results are summarized to identify the function sets represented by the sigmoidal FFANNs with the universal approximation properties. The equicontinuous properties of the identified sets is analyzed. The equicontinuous property is related to the fault tolerance of the sigmoidal FFANNs. The generally used arbitrary weight sigmoidal FFANNs are shown to be nonequicontinuous sets. A class of bounded weight sigmoidal FFANNs is established to be equicontinuous. The fault-tolerance behavior of the networks is analyzed and error bounds for the induced errors established.
Year
DOI
Venue
2004
10.1109/TNN.2004.831198
IEEE Transactions on Neural Networks
Keywords
Field
DocType
universal approximation property,arbitrary weight sigmoidal ffanns,universal approximators,universal approximation result,equicontinuous property,sigmoidal feedforward artificial neural,continuous function,sigmoidal ffanns,fault-tolerance property,bounded weight sigmoidal ffanns,error bound,feedforward sigmoidal network,fault tolerance,equicontinuity,approximation property,fault tolerant,function approximation,artificial neural network
Continuous function,Function approximation,Computer science,Fault tolerance,Artificial intelligence,Artificial neural network,Equicontinuity,Machine learning,Sigmoid function,Feed forward,Bounded function
Journal
Volume
Issue
ISSN
15
6
1045-9227
Citations 
PageRank 
References 
18
1.02
22
Authors
2
Name
Order
Citations
PageRank
Pravin Chandra113514.19
Yogesh Singh226713.87