At this point someone would object that one should use
the most probable values of and , rather than their
expected values. The answer is rather simple.
Let us consider again Eq. ().
Assuming a well precise value of , the probability of Positive
if Infected is exactly equal to .
However, if we want to evaluate
PosInf, taking into account all possible
values of and how much we believe each of them,
that is ,
we just to need to use a well known result of
probability theory:
But, being
PosInf, we get
in which we recognize the expected value of
.21