ABSTRACT
We have seen in Chapter 2 that in attempting to assess the relevant conditional distribution such as pR(u|v) for a referred patient R we are led to the problem of modelling one or other of the conditional distributions pS(u|v) and pS(v|u) for a typical selected case S. Once this modelling problem has been resolved our subsequent task is to use the data set
D = {(ui, vi) : i = 1, . . . , n} of the selected cases S1, . . . ,Sn to fit the model as a step towards arriving at an assessment of the appropriate conditional pS distribution. To fix ideas for the moment we shall concentrate on the modelling and fitting of pS(u|v).
