作者
Min-Rong Chen, Bi-Peng Chen, Guo-Qiang Zeng, Kang-Di Lu, Ping Chu
发表日期
2020/5/28
期刊
Neurocomputing
卷号
391
页码范围
260-272
出版商
Elsevier
简介
The optimal generation of initial connection weight parameters and dynamic updating strategies of connection weights are critical for adjusting the performance of back-propagation (BP) neural networks. This paper presents an adaptive fractional-order BP neural network abbreviated as PEO-FOBP for handwritten digit recognition problems by combining a competitive evolutionary algorithm called population extremal optimization and a fractional-order gradient descent learning mechanism. Population extremal optimization is introduced to optimize a large number of initial connection weight parameters and fractional-order gradient descent learning mechanism is designed to update these connection weight parameters adaptively during the evolutionary process of fractional-order BP neural network. The extensive experimental results for a well-known MNIST handwritten digits dataset have demonstrated that the …
学术搜索中的文章
MR Chen, BP Chen, GQ Zeng, KD Lu, P Chu - Neurocomputing, 2020