在线阅读 --自然科学版 2011年3期《基于遗传算法的RBF神经网络非线性时间序列预测》
基于遗传算法的RBF神经网络非线性时间序列预测--[在线阅读]
郭兰平, 俞建宁, 张建刚, 漆玉娟, 张旭东
兰州交通大学数理与软件工程学院, 甘肃兰州 730070
起止页码: 244--247页
DOI:
摘要
提出一种基于遗传算法和RBF神经网络相结合的时间序列预测模型,克服了单个神经网络在非线性时间序列预测中容易陷入局部极小值及网络训练速度缓慢的问题.以居民消费价格指数数据进行训练和测试,与传统的BP神经网络预测模型相比较,该模型的预测精度是令人满意的,数值模拟证明了该方法的有效性和可行性.

Nonlinear Time Series Forecasting of RBF Neural Network Based on Genetic Algorithm
GUO Lanping, YU Jianning, ZHANG Jiangang, QI Yujuan, ZHANG Xudong
School of Mathematics Physics and Software Engineering, Lanzhou Jiaotong University, Gansu Lanzhou 730070, China
Abstract:
A time series forecasting model based on genetic algorithm and RBF neural network is proposed.The problem that single neural network in nonlinear time series forecasting easily gets into the local minimum and neural network has a very slow study rate is overcome.The model is then used to forecast the inhabitant consumer price index(CPI).Compared with the traditional BP neural network forecast model,this model forecast precision is satisfying.Numerical simulation illustrates the feasibility of the technique.

收稿日期: 2010-6-15
基金项目: 国家自然科学基金(408034);甘肃省自然科学基金(3ZS051-A25-030;3ZS-042-B25-049);兰州交通大学大学生科技创新基金(DXS2010-021)

参考文献: