Comment on “Confidence, credibility and prediction”
- 36 Downloads
My comments focus mainly on the paper’s contribution to prediction which is presented largely through two examples in the second part of the paper.
It is easier to work with prediction of continuous than discrete random variables, so we first consider a slightly simplified version of the second example (Sect. 4.3 in the paper). Suppose we want to predict the value of \(Y|\mu \sim N(\mu ,1)\)
KeywordsPosterior predictive distribution Prediction interval Predictive coverage
© Sapienza Università di Roma 2018