摘要
本文介绍了一种分析模拟模型可靠性的新方法——置信区间法.这种方法根据模型的输出数据和从该模型所对应的实际系统中搜集到的相应数据.在一定置信水平上.利用建立并分析置信区间的方法.对模型的可靠性进行评价.
The paper describes an approach for comparing a model with a system for the situation where it is possible to collect a potentially large amount of data from both the model and the system. Suppose that we have constructed a 100· (1-a)% cofidence interval for ξ(ξ=μx-μy, μx and Wμy are the average of observations for the system data and the model data, respectivly), and let l(a) and u(a) be the corresponding lower and upper confidence intervals end points, respectively. If 0∈(l(a), u(a)), the observed difference between μx and μy might be caused by sampling fluctuation, therefor, we say that the model is valid. If 0∈(l(a), u(a)), the observed diffence between μx and μy is said to be statistically sigificant.
关键词
模拟模型
置信区间
可靠性
statistically significant
practical significant
confidence interval
validation
