摘要
很多预测模型都是利用原始数据直接地代入模型中优化参数,这样仍然很难避免整体数据之间的相互约束,使某些局部误差还是很大.为了克服这些缺点,特提出一种基于伸缩变换不一致性的灰色磨光优化模型,模型通过引入描述局部性质的磨光因子、可逆变换和优化方法,很好地保证整体的拟合精度.通过实例比较,模型比其它几种模型的拟合精度更高.
Many forecast models are used directly to substitute the original data into the model to optimize the parameters of demand, this is still difficult to avoid binding between the overall data, and local error is still very large. In order to overcome these shortcomings, the grey smoothing optimization model basing on inconsistency of stretching transformation is proposed. The model introduces polished factor that describe the partial property, reversible transformation and optimization method, ensuring well the overall fitting accuracy. Compared by exaznple, the fitting accuracy of the model is better than that of several other models.
出处
《数学的实践与认识》
CSCD
北大核心
2010年第18期130-133,共4页
Mathematics in Practice and Theory
基金
江西省教育厅科研项目GJJ09270
关键词
伸缩变换
磨光法
逆变换
原始值参数化
优化
Stretching transformation
smoothing method
inverse transformation
originaldata parameterized
optimization
