摘要: |
基于混沌时间序列的重构相空间、遗传算法的良好全局搜索和神经网络精确的局部搜索特性,以重构相空间中的饱和嵌入维数作为神经网络输入层节点数。通过采用遗传算法优化神经网络初始权重,将重构相空间、遗传算法、神经网络三者有机地结合,提出并建立了相空间遗传BP神经网络预测模型。将该模型用于黄河上游月径流预测。结果表明,该模型应用在水文时间序列的预测中是合理、可行的,并具有较高的精度。 |
关键词: 重构相空间 遗传算法 BP神经网络 径流预测 |
DOI: |
分类号: |
基金项目:国家“973”重点基础研究发展规划项目(G19990436);陕西省重点实验室基金项目(02JS37) |
|
The study on runoff prediction model of BP neural network based on phase space and GA |
|
Abstract: |
Considering the character of reconstruction phase space of chaos time series,the good global searching ability of GA and the effective local searching capacity of BP neural network,the runoff prediction model of BP neural network based on phase space and GA was purposed in this paper,through combining reconstruction phase space,GA with BP neural network.The model is used to predict monthly runoff of up-stream in Yellow River.The result of calculation shows the model is feasible,reasonable and highly precise. |
Key words: reconstruction phase space GA BP neural network runoff prediction |