Performance assessment of the anticipatory approach to optimal experimental design for model discrimination
Performance assessment of the anticipatory approach to optimal experimental design for model discrimination
复制标题
DOI：
10.1016/j.chemolab.2011.06.008
复制
发表时间：
2012-01-15
影响因子：
3.9
通讯作者：
De Baets, Bernard
中科院分区：
文献类型：
--
作者：
Donckels, Brecht M. R.;De Pauw, Dirk J. W.;De Baets, Bernard
当提出几个模型来描述同一个过程时，模型歧视的问题就会出现，这是许多研究领域都会遇到的情况。为了从一组竞争模型中识别出最佳模型，可能需要收集有关过程的新信息，因此必须进行额外的实验。在文献中已经描述了几种方法来设计最佳判别实验。预期的方法是其中之一，是非常有吸引力的概念的角度来看，因为预期的信息内容的新设计的实验被认为是，甚至在实验进行（预期设计）。在本文中，这种方法的性能进行评估，通过比较它与其他的性能，建立的方法，以最佳的实验设计模型歧视。为了进行这种比较，定义了四个性能指标：（1）是否可以识别最合适的模型，（2）必须设计和执行以实现模型区分的额外实验的数量，（3）最终被识别为最合适的模型的参数估计值的质量，以及（4）识别不适当模型的速度。结果清楚地表明，预期的方法有它的好处，并可能是首选的方法在许多应用（生物）化学工程和计算机生物学。(C)2011 Elsevier B.V.保留所有权利。
The problem of model discrimination arises when several models are proposed to describe one and the same process, a situation encountered in many research fields. To identify the best model from the set of rival models, it may be necessary to collect new information about the process, and thus additional experiments have to be performed. Several approaches have been described in literature to design optimal discriminatory experiments. The anticipatory approach is one of them and is very appealing from a conceptual point of view because the expected information content of the newly designed experiment is considered, even before the experiment is performed (anticipatory design). In this paper, the performance of this approach is evaluated by comparing it with the performance of other, established approaches to optimal experimental design for model discrimination. To conduct this comparison four performance measures were defined: (1) whether the most appropriate model could be identified, (2) the number of additional experiments that have to be designed and performed to achieve model discrimination, (3) the quality of the parameter estimates of the model that is eventually identified as the most appropriate one, and (4) the rate at which the inadequate models are identified. The results clearly indicate that the anticipatory approach has its benefits and may be the preferred approach in many applications in (bio)chemical engineering and in-silico biology. (C) 2011 Elsevier B.V. All rights reserved.