Bayesian Diagnostics

在实际计量经济学分析中发生的主要变化之一是越来越多地使用诊断统计，用于批评计量经济学模型并刺激模型修订。诊断统计的使用对统计理论提出了挑战，无论是经典的还是贝叶斯的。传统的统计理论处理的是对假设数据集的计划反应的评估。事实上，如果没有一组表明对每个可能的数据集的数据的响应的计划，就不可能计算抽样属性。使用诊断统计量来批评一个模型是预先宣布计划的响应没有完全承诺，并且可能在观察到实际数据时进行修改。本项目的目的是制定诊断统计。大部分的工作将致力于启发诊断，但批评方面也有一些工作要做。这个方法是贝叶斯的。在为选择变量的问题制定启发式诊断方法方面已经取得了进展。这项工作的扩展将用于处理其他严重的统计问题，如异方差、非正态性和序列相关性。这项工作很重要，因为它为估计经济关系提供了更好的方法。

One of the major changes taking place in practical econometric analysis is the increasing use of diagnostic statistics which are used to critique econometric models and to stimulate model revisions. The use of diagnostic statistics presents a challenge to statistical theory, classical or Bayesian. Traditional statistical theory deals with the evaluation of planned responses to hypothetical data sets. Indeed, it is impossible to compute sampling properties without a set of plans indicating the response to the data for every conceivable data set. The use of a diagnostic statistic to criticize a model is an advance announcement that the planned response is not fully committed and may be revised when the actual data are observed. The purpose of this project is to formulate diagnostic statistics. Most of the effort will be devoted to elicitation diagnostics, but some work well also be done on criticisms. The approach will be Bayesian. Progress has already been made on the formulation of elicitation diagnostics for the problem of choice of variables. Extensions of this work will be undertaken to deal with other serious statistical problems such as heteroscedasticity, nonnormality and serial correlation. This work is important because it will better methods for estimating economic relationships.