The methods used to build estimators are linked with the precise nature of the observations, each kind of observations generating a different statistical problem. We detail these facts in the next sections. We have intentionally omitted in this chapter the additional problem of noisy observations, which often occurs in practice.