We describe how statistical association models and, specifically, graphical
models, can be usefully employed to model web mining data. We describe som
e methodological problems related to the implementation of discrete graphic
al models for web mining data. In particular, we discuss model selection pr
ocedures.