摘要
针对线性组合预测方法的局限性，本文提出了一种基于高斯型模糊逻辑系统的非线性组合预测新方法，并给出了相应的反向传播学习算法确定模糊系统的参数及模糊子集的划分．理论分析和应用实例表明：该方法具有很强的学习与泛化能力，在处理诸如非线性系统中时间序列的组合建模与预测方面都良好的应用价值．
In this paper，a new nonlinear combination forecasting method based on fuzzy logic system ispresented to overcome the limitation in linear combination forecasting． Furthermore，the correspondingback propagation learning algorithm is put forward to identify the parameter of the fuzzy system model and partitions of fuzzy subsets． Theoretical analysis and forecasting examples all show that the new techniqueshas reinforced learning properties and universalized Capabilities．With respect to combined modelingand forecasting of non－stationary time series in nonlinear systems，winch have some uncertainties， themethod are feasible and effective．
出处
《管理科学学报》
1999年第3期28-32,38,共6页
Journal of Management Sciences in China
基金
国家自然科学基金!79770105
关键词
组合预测
模糊逻辑系统
反向传播
学习算法
combination forecasting，fuzzy logic system，back propagation learning algorithm
