In the simplified NLIN example below, giving initial values to model parameters in the PARAMETER statement is NOT equivalent to specifying them in a dataset. It looks like parameter names are not known at the time of array definitions when the parameters are given in a dataset. Is this an undocumented "feature"?
/* Parameter array given initial values in PARAMETER statement. Works fine. */
proc nlin data=sashelp.class;
parameters w=1 h=1;
bounds w h >= 0;
array x{*} weight height;
array y{*} w h;
s = 0;
do i = 1 to dim(x);
s = s + x{i}*y{i};
end;
model age = s;
run;
/* Same problem. Parameter array given initial values in PDATA dataset. Doesn't work. */
data pd;
Parameter="w"; Estimate=1; output;
Parameter="h"; Estimate=1; output;
run;
proc nlin data=sashelp.class;
parameters / pdata=pd;
bounds w h >= 0;
array x{*} weight height;
array y{*} w h;
s = 0;
do i = 1 to dim(x);
s = s + x{i}*y{i};
end;
model age = s;
run;
ERROR: The variable y was referenced but not given a value.
Available on demand!
Missed SAS Innovate Las Vegas? Watch all the action for free! View the keynotes, general sessions and 22 breakouts on demand.
ANOVA, or Analysis Of Variance, is used to compare the averages or means of two or more populations to better understand how they differ. Watch this tutorial for more.
Find more tutorials on the SAS Users YouTube channel.