In the first example, the assignment named A is defined to be the
linear combination . None of these quantities need exist
at this point, and if they do not exist when the assignment is used,
they will be ignored. The examples include such quantities as %c
(meaning the constant c) and #kda.s (meaning the
function kda.s), which are described more fully in
chapter 7.
BD>assign : G.3.4=(%c)+H.2+F+ (26)
BD>assign : D.s.t=(#kda.s+#kdb.s) D.s.(t-1) + (#kdb.s)
D.s.(t-2)+(#kdc.s)+E.s.t
In the second of these examples, the assignment named is
defined to be the linear combination
, where
%c is some constant which will be evaluated at this point (use
a function rather than a constant if you wish to delay the evaluation). Notice
that this example contains two scalar parts.
The third example is taken from a genuine application, and involves part of the specification of an autoregressive error structure. It makes the definition
where the 's are functions of s, the D's are vectors of
quantities related to similar previous quantities via the given
relationship, and the
terms are noises.
The fourth example shows a definition part which consists of a component
which includes varying indices, and a component which does not. Notice
that this results for example in the definition of
, so that the component
is repeated.
This is permissible, and the definition part will be taken to be
.