I'm using spikeslab for the first time, but don't understand what the output means. It was suggested to me that I use it to tell which variables my dependent variable is most correlated to, in a ranked order. Particuarly, what is "bma" bma.scale" "gnet" and "gnet.scale"? I also don't understand how to read the corresponding plot to the model.Thanks for any help!

For example, this is one of the models I created using spikeslab, with its output:

model2_ss <-spikeslab(Risk_Pct ~ Race
+                    +hispanic
+                    +Born_In_US
+                    +Highest_Education
+                    +Marital_Status
+                    , na.rm = TRUE, data = LabeledData)
> model2_ss
Variable selection method     : AIC 
Big p small n                 : FALSE 
Screen variables              : FALSE 
Fast processing               : TRUE 
Sample size                   : 26 
No. predictors                : 5 
No. burn-in values            : 500 
No. sampled values            : 500 
Estimated mse                 : 0.4299 
Model size                    : 3 

---> Top variables:
                bma      gnet     bma.scale     gnet.scale
Marital_Status  0.516     0.516     0.319         0.319
Born_In_US      -0.469    -0.447    -0.440       -0.419
Race            0.458     0.421     0.926         0.851

Plot of model2_ss


BMA is "Bayesian Model Averaged". GNET is "Generalized Elastic Net".

Have you tried reading the Ishwaran and Rao papers as mentioned in the documentation for spikeslab? There's an article in the R Journal as well that might be worth reading too: https://journal.r-project.org/archive/2010-2/RJournal_2010-2_Ishwaran~et~al.pdf - no sense duplicating it all here.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.