[pymvpa] Regression estimates?

[Per B. Sederberg]

Hi folks:

Long time no chat  :)

I've been struggling for hours to do something extremely easy (with
latest github master), so I'm breaking down to ask.  I'm running a
cross-validated regression and sensitivity analysis:

clfr = GLMNET_R(alpha=.5, model_type='naive', enable_ca=['estimates'])
sclf = SplitClassifier(clfr,
                       NFoldPartitioner(),
                       enable_ca=['stats','estimates'])

# Compute sensitivity, which internally trains the classifier
print "Running CV"
analyzer = sclf.get_sensitivity_analyzer()
senses = analyzer(dat_all)
print sclf.ca.stats

I'm having trouble understanding the performance in the stats:

-------> print(sclf.ca.stats)
Statistics  Mean  Std   Min      Max
---------- ----- ----- -----    -----
Data:
   RMP_t   2.285 0.696  1.3       4
   STD_t     0     0     0        0
   RMP_p   2.357 0.627 1.494    4.502
   STD_p     0     0     0        0
Results:
    CCe      1     0     1        1
   RMSE    0.392 0.267 0.019    1.117
RMSE/RMP_t  0.18 0.125 0.009    0.477
Summary:
    CCe     0.25         p=  0.000125809
   RMSE     0.47
RMSE/RMP_t  0.2
 # of sets   20


Also, I really just want to get the predictions/estimates of the
classifier from each fold, but I don't know how to get that out.
Every place I look for estimates just gives me:

UnknownStateError: Unknown yet value of estimates


Any pointers would be greatly appreciated as I have about one more
hour to get this working.

Thanks,
Per