Show simple item record

dc.contributor.author林耀进
dc.contributor.author吴顺祥
dc.date.accessioned2016-05-17T02:42:45Z
dc.date.available2016-05-17T02:42:45Z
dc.date.issued2009
dc.identifier.citation计算机工程与应用,2009,(18):238-239+242
dc.identifier.issn1002-8331
dc.identifier.otherJSGG200918072
dc.identifier.urihttps://dspace.xmu.edu.cn/handle/2288/105120
dc.description.abstract在分析gM(1,1)模型的建模机理的基础上,指出了传统建模方法的不足,即发现了预测数据序列中的第一点的值并不能用原始数据序列中第一点的值来代替,因为存在误差,同时给出了误差项的一般表达式,然后基于bP神经网络对误差项进行优化模型。结果表明,该模型拟合误差小,预测精度高。
dc.description.abstractAccording to the building mechanism of GM(1,1),the existent weakness is pointed out to of traditional method build grey model,that is,the first point of original data is different with 1st point of predictive value that both exist an error term μ.Furthermore,by the error term μ,this paper formulates other error term of training data.The optimum model is given to error term on the basis of BP neural network.The results of experiment show that the model is valid,feasible and high precision.
dc.description.sponsorship国家自然科学基金No.60704042;国家"十一五"科技支撑计划项目No.2007BAK34B04;厦门大学985二期信息创新平台项目----
dc.language.isozh_CN
dc.subject灰色系统
dc.subjectBP神经网络
dc.subjectGM(1
dc.subject1)模型
dc.subject误差项
dc.subjectgrey system
dc.subjectBP neural network
dc.subjectGM(1
dc.subject1) model
dc.subjecterror term
dc.title灰色误差神经网络模型在预测中的应用研究
dc.title.alternativeApplication research of grey error term and neural network model on prediction
dc.typeArticle


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record