一种改进的模糊变权重组合预测方法

来源期刊：中南大学学报(自然科学版)2003年第6期

论文作者：李学全 李春生

文章页码：708 - 710

关键词：FVW;模糊控制;组合预测;灰色关联度

Key words：Fuzzy Variable Weight; fuzzy control; combination forecast; grey interconnect degree

摘    要：在对FVW方法的实质进行分析的基础上,根据预测对象的变化趋势对加权系数的影响,利用预测相对误差、预测对象的变化趋势,对曹长修等提出的FVW方法加以改进,克服了原有FVW方法中用预测绝对误差和预测绝对误差的变化率为模糊控制因素时的因素相关性;通过灰色关联度引入基本权重概念,以刻画各种方法在某时刻前一段时期内全面的、平均的预测效果对加权系数的影响.仿真结果表明,采用改进后的FVW法,预测平均误差和预测平方根误差均较小.

Abstract: Based on the analysis of the essence of Fuzzy Variable Weight (FVW) combination forecasting, an improve FVW method was proposed by making use of the relative error and the tendency of the forecasted object. The improve FVW method overcame the correlation between fuzzy control factors, the absolute error and the ratio of absolute error.Through the concept of gray interconnect degree, the basic weight to depict the influence of mean prediction effect befor some moment on the weights was studied. The simulation results show that the improved method of FVW is more effective.