Petersen B J
Novigen Sciences, Inc., Washington, DC 20036, USA.
Food Addit Contam. 2000 Jul;17(7):591-9. doi: 10.1080/026520300412500.
Probabilistic modelling techniques allow much more realistic estimates of exposure and risk by computing the use of the full range of potential exposures rather than single 'worst case' exposures. However, these techniques require additional considerations regarding the appropriate data and models. This article reviews the theoretical aspects of probabilistic modelling and also considers some of the practical applications. The most common method, called Monte Carlo analysis, is discussed in some detail. The practical application of Monte Carlo to risk assessments is presented along with an evaluation of the input parameters. Topics also discussed include considerations of the requirements for precision and procedures for validation of assessments.
概率建模技术通过计算所有潜在暴露范围的使用情况，而非单一的“最坏情况”暴露，能够对暴露和风险进行更为现实的估计。然而，这些技术在适当的数据和模型方面需要额外考虑。本文回顾了概率建模的理论方面，并考虑了一些实际应用。文中详细讨论了最常用的方法——蒙特卡罗分析。介绍了蒙特卡罗在风险评估中的实际应用以及对输入参数的评估。还讨论的主题包括对精度要求的考量以及评估验证程序。
