摘要
预报因子选择与模型训练精度确定,是模糊优选神经网络模型应用于径流中长期预报时有待研究解决的两个重要问题.应用预报因子集与预报量间的复合非线性相关分析方法选择预报因子(集),克服了通常单因子线性相关分析选择预报因子的不适用性;通过定义综合效应系数来综合评价模糊优选神经网络模型的拟合能力与外推预报能力,为研究模型的拟合精度高而外推预测精度低的问题提供了一种解决方法.
Medium-term and long-term runoff forecast is significant to reservoir operation and water resources utilization. In order to improve the lower level of medium-term and long-term runoff forecast, a fuzzy optimization neural network model is imported. A primary issue is how to choose factors since the common linear coefficient is not adaptable for the nonlinear system. Another issue is that higher level of fitting precision and lower level of forecasting precision coexist. By defining a nonlinear coefficient, the complex nonlinear relations are analyzed between forecast factors and forecast object. The effective factors are chosen according to the nonlinear coefficient. Integrative effective coefficient is defined to evaluate the effect of forecast model. It provides a solution for the problem, higher level of fitting precision and lower level of forecasting precision.
出处
《大连理工大学学报》
EI
CAS
CSCD
北大核心
2008年第3期411-416,共6页
Journal of Dalian University of Technology
基金
国家自然科学基金委员会
二滩水电开发有限公司雅砻江水电联合研究基金资助项目(50579095)
关键词
模糊优选
神经网络
年径流预报
预报因子选择
综合效应系数
fuzzy optimization
neural network
annual runoff forecast
selection of forecast factors
integrative effective coefficient
