The lecture gives an introduction to Bayesian inference and describes how graphical models can be extended to include parameters explicly, thus building a bridge between Bayesian networks and traditional statistical modelling, such as linear regression analysis etc.
The lecture is largely based on
D. J. Spiegelhalter (1998). Bayesian graphical modelling: a case-study in monitoring health outcomes. Applied Statistics, 47, 115-133.
Overheads for this and next lecture for screen-viewing and printing