Dear Experts,
I just finished my BMS and found the best model. But I have to get all the parameters for the optimal model when I start to write. I have two questions for getting these parameters.
1) I know I can use the new BMA function to get these parameters, however when I used the function, there is an error following like this:
Running 'BMS: DCM'
Loading model space
Computing FFX model/family posteriors ...
FFX Bayesian model averaging ...
6 models in Occams window
Model 15, p(m|Y)=0.13
Model 16, p(m|Y)=0.14
Model 17, p(m|Y)=0.01
Model 18, p(m|Y)=0.07
Model 19, p(m|Y)=0.07
Failed 'BMS: DCM'
Attempt to reference field of non-structure array.
In file "C:\Program Files\MATLAB71\toolbox\spm8\spm_dcm_bma.m" (v3966), function "spm_dcm_bma" at line 255.
In file "C:\Program Files\MATLAB71\toolbox\spm8\config\spm_run_bms_dcm.m" (v3955), function "spm_run_bms_dcm" at line 341.
The following modules did not run:
Failed: BMS: DCM
The syntax at line 255 is: dimd = size(params(1).model(1).Ep.D,3);
I checked and found no data called params(1).model(1).Ep.D, what is D here representing for? I don't know why the BMA fails, but the BMS ffx works very well.
Thanks
2) as I failed in BMA, then I have to use the classical t-test approach. I loaded the best model's DCM file in every participant's folder. However, I was frustrated about the parameters called A, pA, vA. What are the differences between them? Which one should I use for the intrinsic connection of the best model?
Thank you very much
Yin
--
Yin Wang
Antonia Hamilton's Social Cognition Lab
School of Psychology
University of Nottingham,
University Park, Nottingham, UK
NG7 2RD
[log in to unmask]