On the Approximate Elimination of Nuisance Parameters by Conditioning
研究了在存在干扰参数时，如何通过条件化方法对感兴趣的标量参数进行最优推断，并分析了信息损失与模型统计曲率的关系。
The general problem of inference about a scalar parameter of interest θ in the presence of a nuisance parameter λ using conditional inference is considered.xml . A condition is given under which inference based on the conditional distribution of θ^, the maximum likelihood estimate of θ, given λ^o, the maximum likelihood estimate of λ for fixed θ = θ0, is optimal, in a certain sense. When this condition is not satisfied, it is shown that inference should be based on the conditional distribution of θ^ given λ^o, j^o where j^o denotes the observed information for λ for fixed θ =θθo, although this will involve some loss of information about θ. This information loss is shown to be related to the statistical curvature of the model.