I'm doing a sanity check with my auditory fMRI task and am trying to get 3dsvm to classify sound=1 vs no sound=0.
when running this script:
3dsvm -testvol run6_detrended+tlrc. -model model_runs1thru5+tlrc. -testlabels 36289_MVPArun6.1D -predictions pred6_model1thru5
the predictions that are spit out are not zeros and ones, but fractions (e.g.
0.042139769
0.85066682
0.79669857
0.57938343
1.1055593
0.78475887
0.077376604
.
.
.)
I double checked my training text files to make sure they were not the source of the problem, then added -classout to the above script. This helped, except, not only was I getting 1s and 0s, but also -0s. Anyone know why this would be, and does the classifier count -0 as 0?
Thank you!
Emily