摘要
GM(1,1)模型的实质是小样本、贫信息下的预测模型,其目的是得到误差尽可能小的预测值.在分析GM(1,1)模型建模机理的基础上,提出了GM(1,1)模型中参数a,b的一种新算法——神经网络算法.把神经网络中的BP算法应用于GM(1,1)模型的建模过程,实例表明可使预测精度得到提高.
The essential of GM(1,1) model is a kind of model to forecast under little sample and little information. The aim of the model is to get forecasted value with little error as soon as possible. Based on analyzing the mechanism of GM(1,1) model, this article puts forward a new method of parameter a, b in GM(1,1) model, namely, Neural Network calculating. BP calculating method is applied into the modeling progress of GM (1,1) model, and Examples suggest that the forecasting precision of the improved model gets better result.
出处
《数学的实践与认识》
CSCD
北大核心
2006年第4期126-130,共5页
Mathematics in Practice and Theory
基金
江苏省自然科学基金(BK99109)
江苏省教育厅高校自然科学研究计划(4KJD130039)