摘要
考虑到某月径流与该月历史同期径流以及临近月径流均有较强相关性,而通常预报方法只采用其中一种径流序列,导致了可用信息损失。为此,提出一种基于灰色理论和支持向量机回归的组合预报模型。提出的模型综合利用了径流年内变化和年际变化信息,与单一灰色模型和支持向量机模型进行预测对比,结果表明基于灰色支持向量机的月径流模型预测精度明显高于单一模型,尤其是对径流变化剧烈的汛期表现出更优越的预测性能。
The monthly runoff has a strong correlation with both the same period runoff in the history and nearby runoff. The previous prediction methods for monthly runoff often use only one kind of runoff series, which led to the loss of available information. So, a combination forecasting model based on Grey Theory and Support Vector Machine regression is presented. The model comprehensively utilizes annual and inter-annual runoff information. Compared to single grey model and support vector machine model, the results show that the performance of combination model is obviously better, especially for the flood season runoff predictions.
出处
《水力发电》
北大核心
2015年第12期17-20,共4页
Water Power
基金
国家自然科学基金项目(51239004)
高等学校博士学科点专线科研基金资助项目(20100142110012)
关键词
支持向量机
灰色理论
核函数
月径流
Support Vector Machine
Gray Theory
Kernel Function
monthly runoff
