Title :
Evolving recurrent neural networks with non-binary encoding
Author :
Mandischer, Martin
Author_Institution :
Syst. Anal. Res. Group, Dortmund Univ., Germany
fDate :
29 Nov-1 Dec 1995
Abstract :
This paper presents an evolutionary approach for the design of feedforward and recurrent neural networks. We show that evolutionary algorithms can be used for the construction of networks for real-world tasks. Therefore, a data structure based genotypic network representation, as well as genetic operators, are introduced. Results from the classification, function approximation and time-series domains are presented
Keywords :
data structures; encoding; feedforward neural nets; function approximation; genetic algorithms; pattern classification; recurrent neural nets; time series; classification; data structure based genotypic network representation; evolutionary algorithms; evolutionary design approach; feedforward neural networks; function approximation; genetic operators; nonbinary encoding; real-world tasks; recurrent neural networks; time series; Computer science; Electronic mail; Encoding; Evolutionary computation; Feedforward systems; Function approximation; Genetic algorithms; Network topology; Neural networks; Recurrent neural networks;
Conference_Titel :
Evolutionary Computation, 1995., IEEE International Conference on
Conference_Location :
Perth, WA
Print_ISBN :
0-7803-2759-4
DOI :
10.1109/ICEC.1995.487449
