首页> 外文会议>IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology >Determining structural parameter identifiability in biological dynamical models by analysing the statistical properties of the likelihood behaviour
【24h】

Determining structural parameter identifiability in biological dynamical models by analysing the statistical properties of the likelihood behaviour

机译:通过分析似然行为的统计特性确定生物动力学模型中的结构参数可识别性

获取原文

摘要

Identifiability is a fundamental precondition for model identification and concerns the possibility to determine unique parameters sets using a certain model structure and the experimental observations for the model input-output. In this article we show how to a priori assess the identifiability of model parameters analysing the properties of the likelihood function. More precisely, a model is identifiable if the likelihood function shows a unique absolute minimum beyond the confidence intervals. The study of the likelihood function cannot avoid accompanying the likelihood estimates with the assessment of the confidence intervals, which reflect the variability in experimental data. We then show the results of this approach in cases of biotechnological relevance where the identifiability has been previously assessed.
机译:可识别性是模型识别的基本前提条件,并且涉及使用某种模型结构确定唯一参数集的可能性以及模型输入输出的实验观察。在本文中,我们展示了如何先验评估分析似然函数属性的模型参数的可识别性。更确切地说,如果似然函数显示超出置信区间的唯一绝对最小值,则可以识别模型。对似然函数的研究无法避免随着置信区间评估的可能性估计,这反映了实验数据的变化。然后,在先前评估可识别性的生物技术相关性的情况下,我们展示了这种方法的结果。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号