Noninformative Priors for Inferences in Exponential Regression Models
研究了指数回归模型中回归参数的推断难题，提出参考先验方法能给出满意的推断结果，并通过具体例子展示了估计和可信区间。
In the exponential regression model, inference concerning the regression parameter is notoriously difficult, even when using the Bayesian noninformative prior approach. The reference prior approach (Bernardo, 1979; Berger & Bernardo, 1989) is considered, and argued to yield very satisfactory inferences. Estimation and credible sets are considered in a specific example.