Dear Stewart:
The only fields that should differ are the DCM.pA, pB, and pC fields. That
is because when the DCM is calculated the probabilities are stored using a
threshold of log(2)/4, while with displaying DCMs you may have used a
threshold of 0 or other threshold. You can check this by typing log(2)/4 at
the threshold prompt when loading a DCM.
Darren
At 03:06 PM 12/14/2004, Stewart Denslow wrote:
>I have noticed some disagreements between two modes of output within the DCM
>module. After specifying and estimating a DCM, one can review the aspects
>of the model with graphics output. Upon quitting from this option, the
>program apparently outputs a structure to "ans" in the matlab command
>window. This allowed me to conveniently just type in DCM=ans to get the
>full structure which I then examined and copied at my leisure. To my
>dismay, I later found that the values in the data structure obtained this
>way were not always the same as what is obtained from loading the
>DCM_???.mat file. Which one of these is correct? I would guess that the
>values in the DCM.mat file are correct but why are the others different?
>
>Can anybody explain this mess?
>
>Thanks,
>Stew Denslow
-------------------------------------------------------------------------
Darren R. Gitelman, M.D.
Cognitive Neurology and Alzheimer¹s Disease Center
Northwestern Univ., 320 E. Superior St., Searle 11-470, Chicago, IL 60611
Voice: (312) 908-9023 Fax: (312) 908-8789
-------------------------------------------------------------------------
|