Estimators and hypothesis testing

Very often, the information the engineer has to process concerns a random variable, for which he or she has no theoretical element enabling him or her to determine a priori – by reasoning alone – the mathematical expectation or standard deviation.

From then on, he will have to rely on the notion of estimator, built from necessarily partial test results, and on the formulation of "hypotheses" that statistical analysis will lead him to reject or not. All the decisions he makes using these tools will be accompanied by a clearly quantified – consenti –risk.