An improved extreme learning machine based on particle swarm optimization

Authors:
Fei Han;Hai-Fen Yao;Qing-Hua Ling
Affiliations:
School of Computer Science and Telecommunication Engineering, Jiangsu University, Zhenjiang, Jiangsu, China;School of Computer Science and Telecommunication Engineering, Jiangsu University, Zhenjiang, Jiangsu, China;School of Computer Science and Engineering, Jiangsu University of Science and Technology, Zhenjiang, Jiangsu, China
Venue:
ICIC'11 Proceedings of the 7th international conference on Intelligent Computing: bio-inspired computing and applications
Year:
2011

Citing 5
Cited 1

Rapid and brief communication: Evolutionary extreme learning machine

Pattern Recognition
On improving the conditioning of extreme learning machine: a linear case

ICICS'09 Proceedings of the 7th international conference on Information, communications and signal processing
OP-ELM: optimally pruned extreme learning machine

IEEE Transactions on Neural Networks
Evolutionary extreme learning machine – based on particle swarm optimization

ISNN'06 Proceedings of the Third international conference on Advances in Neural Networks - Volume Part I
The sample complexity of pattern classification with neural networks: the size of the weights is more important than the size of the network

IEEE Transactions on Information Theory

Investigating the use of alternative topologies on performance of the PSO-ELM

Neurocomputing

Quantified Score

Hi-index	0.00

Visualization

Abstract

Traditional extreme learning machine (ELM) may require high number of hidden neurons and lead to ill-condition problem due to the random determination of the input weights and hidden biases. In this paper, we use a modified particle swarm optimization (PSO) algorithm to select the input weights and hidden biases of single-hidden-layer feedforward neural networks (SLFN) and Moore---Penrose (MP) generalized inverse to analytically determine the output weights. The modified PSO optimizes the input weights and hidden biases according to not only the root mean squared error on validation set but also the norm of the output weights. The proposed algorithm has better generalization performance than other ELMs and its conditioning is also improved.