摘要
针对传统的神经网络训练算法收敛速度慢和泛化性能低的缺陷,提出一种新的自适应粒子群优化算法用于神经网络的训练.该算法通过改进自适应搜索策略以提高网络泛化性能,并结合Ionosphere雷达信号分类数据集进行仿真测试.研究结果表明:基于自适应粒子群优化算法训练的神经网络在分类准确率和分类误差上明显优于传统的BP算法,且很好地提高了网络泛化能力和优化效果.
In view of the traditional neural network training algorithm defects of slow convergence speed and the low generalization,a new adaptive particle swarm optimization is put out to apply to the neural network training.The algorithm enhanced the generalization of network through improving the self-adapted search strategy.Then,the algorithm used standard data sets,Ionosphere,in the test.The experiment show that the neural network based on adaptive particle swarm optimization is obviously superior to BP algorithm in the classification accuracy rate and sorting error,and enhances the generalization and the optimized effect of the network.
出处
《江西师范大学学报（自然科学版）》
CAS
北大核心
2010年第6期632-635,共4页
Journal of Jiangxi Normal University(Natural Science Edition)
基金
国家自然科学基金(60802030)资助项目
关键词
神经网络
自适应粒子群优化算法
仿真
neural network
adaptive particle swarm optimization
emulation
