
THEORETICAL RESULTS ON ELASTIC PREDICTIVE RANDOM SETS 89
An important application of Theorem 5.1 is when an IM has been created using
a predictive random set satisfying the efficiency criteria in Chapter 4 without con-
sidering constraints. That predictive random set can be used for S
0
when creating
an elastic predictive random set. Thus, when constraints are incorporated using the
elastic method, bel
x
will be valid for inference about any A ⊂C.
5.4.1 Dempster’s rule of combination
Another method for incorporating parameter constraints into an IM is the condition-
ing rule described in [63] known as “Dempster’s rule of combination” or “Demp-
ster’s conditioning ...