ABSTRACT
To begin this chapter, let us repeat one more time that a test procedure (in this case a goodness-of-fit test) only makes it possible to reject or not reject the null hypothesis H0 : F(x) = F0(x). A goodness-of-fit test cannot be applied to choose the best frequency models among F1(x), F2(x), etc. In addition, the validation of the calibration, as discussed in Section 4.1, does not constitute a validation of the model itself; it serves only as an internal validation of the model.
