Normal Models

As in every subsequent chapter, we start with a description of the data used for the whole chapter as a benchmark for illustrating new methods and for testing assimilation of the techniques. We then propose a corresponding statistical model centered on the normal N (µ, σ²) distribution and consider specific inferential questions to address at this level, namely parameter estimation, one-sided test, prediction, and outlier detection, after we set the description of the Bayesian resolution of inferential problems. This being the first chapter, the amount of technical/theoretical material may be a little overwhelming at times. It is, however, necessary to go through these preliminaries before getting to more advanced topics with a minimal number of casualties!