A New Ridgelet Neural Network Training Algorithm Based on Improved Particle Swarm Optimization
Autor: | Rijian Su, Li Kong, Pu Zhang, Kaibo Zhou, Shengli Song, Jingjing Cheng |
---|---|
Rok vydání: | 2007 |
Předmět: |
Meta-optimization
Artificial neural network Computer science business.industry ComputingMethodologies_MISCELLANEOUS Computer Science::Neural and Evolutionary Computation MathematicsofComputing_NUMERICALANALYSIS Stability (learning theory) Particle swarm optimization Recurrent neural network Artificial intelligence Multi-swarm optimization business Algorithm Metaheuristic |
Zdroj: | ICNC (3) |
Popis: | An improved particle swarm optimization is used to train ridgelet neural network instead of the traditional gradient algorithms. Firstly, the model of ridgelet neural network and the traditional particle swarm optimization (PSO) algorithm are briefly described. Secondly, an improved particle swarm optimization with self-adaptation mutation factor is proposed. Then the improved particle swarm optimization is applied to rigdelet neural network training. Experimental results demonstrate that the new algorithm is better than the traditional particle swarm optimization algorithm in training ridgelet neural network. It has both a better stability and a steady convergence, and is easy to be realized. |
Databáze: | OpenAIRE |
Externí odkaz: |