摘要
人工神经网络的结构设计没有系统的规律可遵循,而常用的基于梯度的神经网络参数优化又易陷入局部最优解。针对BP人工神经网络所存在的缺陷,结合差异演化算法,提出了实数编码的DE-BP神经网络预测模型。利用税收预测的实例验证了算法的有效性,取得了令人满意的结果。
Structural designing of artificial network is always a trouble problem without systematic rule and local minimum usually connects with conventional grads based on parameters optimization. Aiming at the drawback in classical BP artificial networks and combining with differential evolution algorithms, this paper puts forwards the prediction model based on real number coded DE-BP artificial networks. This model achieves satisfactory results on tax forecasting,
出处
《计算机工程与应用》
CSCD
北大核心
2008年第14期246-248,共3页
Computer Engineering and Applications
基金
天津市科技支撑计划项目(No.08ZCKFGX01300)
中国民航大学科研启动基金(No.06qd02x)
关键词
差异演化算法
神经网络
预测
实数编码
Differential Evolution ( DE )
neural network
prediction
real-coded