statistics - SAS stdi output clarification -
i'm attempting convert old sas macros python, , bit unclear of terminology used in sas. in macro, proc statement is
proc reg data=model_file; model &y = &x; output out=&outfile r=resid stdi=resid_error;
i understand r
means individual residual per data point, unclear stdi
meant. according sas manual, stdi means "standard error of individual predicted value", there 1 stdi each row in dataset. searched around bit , found lecture slide university of wisconsin believe explains how calculate stdi
:
however, 2 (edit: one) questions remain:
is method calculating standard error of individual prediction in lecture slide indeed correct? i've never seen method before still have doubts. i've looked sas manual, definition stdi bit confusing:. specifically, h(i) defined don't know bar after [x'x] supposed mean.
the way standard error of individual predictions calculated here utilizes
x
. however, happens if have run regression multiple x columns? stdi assume single x column?
answer: answer no. can have multiple x columns , still stdi value.
i'm not statistician , question have included lot more detail, quick google search suggests you're looking @ proc reg. main documentation proc reg here:
and there dedicated page "model fit , diagnostic statistics" including relevant formulae here:
maybe answer question. although these things don't interest me directly, believe sas's documentation pretty @ describing exact computations being done each procedure.
Comments
Post a Comment