Anotation:
Divergence-based test of goodness of fit when the observations are not identically distributed is introduced in this work.Its asymptotic properties are studied and its bootstrap versions are considered. Applications to generalized linear models diagnostic are proposed.