On Parameter Redundancy in Curve Fitting of Kinetic Data

On Parameter Redundancy in Curve Fitting of Kinetic Data
复制标题

动力学数据曲线拟合中的参数冗余问题

DOI:
10.1007/978-1-4613-3255-8_3
复制
发表时间:
1981
期刊:
影响因子:
3.4
通讯作者:
J. Reich
J. Reich
中科院分区:
材料科学3区
文献类型:
--
作者:
J. Reich

文献摘要

被引文献

相似文献

当所提出的动力学模型对于可测量数据的实际信息内容过于详细时,会出现参数冗余。然后,一组完全不同但相互依赖的参数值能够解释数据,参数估计变得不可能。本文表明,在实验可用之前,可以对该缺陷进行研究。数值标准概述,可用于估计冗余和冗余参数组合,预测估计参数的标准偏差,以优化实验设计,减少冗余。
Parameter redundancy arises when the proposed kinetic model is too detailed for the actual information content of the measurable data. Then an enormous set of totally different, but interdependent parameter values is able to explain the data, and parameter estimation becomes impossible. The paper shows that the defect may be studied before experiments are available. Numerical criteria are outlined which can be used to estimate redundancy and redundant parameter combinations, to predict standard deviations of estimated parameters, to optimize the experimental design for reducing redundancy.