摘要
为了确定滞时、嵌入维数和最邻近点数这3个混沌时间序列局域预测模型参数,首先利用关联积分法确定滞时和嵌入维数,重构混沌时间序列的相空间;而后在此基础上,提出了一种新的预测模型——加权动态局域预测模型.该模型综合考虑了广义自由度和邻近点权重,给出了确定最优邻域的判定指标.实际水文系统的计算分析表明,加权动态局域预测模型具有较高的预测精度,是一种有效的用于混沌水文时间序列的预测模型.
The prediction of chaotic time series based on the reconstructed phase space requires the knowledge of three parameters, the time delay, the embedding dimension and the number of the nearest neighbors. The time delay and the embedding dimension are calculated by the correlation integral approach. The weight-dynamic local prediction model is presented based on the reconstructed phase space, which takes the generalized degrees of freedom and neighbors' weight into account, and the decisive condition of the optimal neighborhood is proposed. The results indicate that the proposed model has superior predictive capability, and is a valid prediction model for chaotic hydrological time series.
出处
《大连理工大学学报》
EI
CAS
CSCD
北大核心
2004年第3期445-448,共4页
Journal of Dalian University of Technology
