关键词: 神经网络, 群智能算法, 全局最优解, 优异性能

Abstract:

It is not reasonable to select the appropriate weights to bring out the desired results with the global optimal solution on the misleading of local optimal solution in the process of neural network model.A swarm intelligence algorithm is used to obtain the global optimal solution and the neural network model is established by using the weight and threshold value of the sample to the lowest point. The results showed that the evaluation parameters of the network model showed good performance, the prediction accuracy and the correlation coefficient were 97.55% and 96.2% respectively among them. Examples show that the neural network based on swarm intelligence algorithm can be able to get the global optimal solution and the model has good performance under the modeling condition which can prove its great value in theory and market.

Key words: neural network, swarm intelligence algorithm, global optimal, modeling conditions