Normal means revisited
The authors suggest two criteria for ranking and selecting which provide insight not only to which population is best, but in addition give quantitative methods for deciding how much better one population is than another. \textsl{Criterion 1: The posterior probability that any one of the means is larger than all of the others by an amount ``\(b\), i.e., compute for each \(i=1,2,\ldots,k\), the quantity \[ P_{i}=P(\nu_{i}\geq\nu_{j}+b \;\text{for all} \;j\neq i|\text{data}), \] where \(b\) is a non-negative specified constant;} \textsl{Criterion 2: The predictive distribution that any one of these populations will have a larger observation than all the others by an amount ``\(c\), i.e., compute for each \(i=1,2,\ldots,k\), the quantity \[ PR_{i}=P(Y_{i}\geq Y_{j}+c \;\text{for all} \;j\neq i|\text{data}), \] where \(Y_{i}\) is a new observation from the \(i\)-th population and \(c\) is a non-negative specified constant.} The authors consider these criteria for ranking of sample means \(\theta_{i}\) from normally distributed populations. They suppose that \(\theta_{i}\), \(i=1,\ldots,k\), are conditionally i.i.d. with a normal distribution with mean \(\beta\) and variance \(\tau^2.\) For the entire collection see [Zbl 0990.62506].
- On the performance of subset selection rules under normality
- On the performance of Gupta's subset selection rule
- scientific article; zbMATH DE number 4134006 (Why is no real title available?)
- Bayesian subset selection approach to ranking normal means
- Complete ranking procedures with quadratic loss
- A Bayesian multiple comparison procedure for ranking the means of normally distributed data
This page was built for publication: Normal means revisited
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3843158)
