摘要
提出综合利用系统前向预测和后向预测对混沌时间序列进行建模,从物理学原理上解释了该模型相对于前向预测模型和后向预测模型的好处.计算机仿真结果表明:对于时间可逆性较好的混沌系统,前后向联合预测模型的建模性能比前向预测模型好,前向预测模型比后向预测模型预测性能好,但对于时间可逆性差的混沌系统,前后向联合预测模型的建模性能较前向预测模型略差,而后向预测模型比前向预测模型的预测性能差了很多.
A forward backward united prediction model based on radial basis function neural networks for chaotic time series is proposed, and it utilizes forward prediction and backward prediction simultaneously. The reason why the proposed model has better performance than the forward prediction model and the backward prediction model is given. The results of computer simulation show that, for the chaotic systems with good time reversibility, the proposed model has better modeling performance than the forward prediction model, which has better modeling performance than the backward prediction model. For the chaotic systems with bad time reversibility, the proposed model has a little worse modeling performance than the forward prediction model, and the backward prediction model has much worse modeling performance than the forward prediction model.
出处
《武汉理工大学学报（交通科学与工程版）》
2007年第2期259-261,340,共4页
Journal of Wuhan University of Technology（Transportation Science & Engineering）
基金
国家重点基础研究项目(批准号:5132102ZZT32)
国家重点实验室基金项目(批准号:514450801JB1101)资助
