Empty gibes_res

[wangtaifr]

Hello. I tested GSimp with my own dataset using parallel mode. Imputations went fine except I noticed the gibbs_res slot was empty:

str (data.imputed$gibbs_res)
 logi[1:3, 0 , 1:2000] 
 - attr(*, "dimnames")=List of 3
  ..$ : chr [1:3] "std" "yhat" "yres"
  ..$ : NULL
  ..$ : NULL

I wonder if there's anything wrong within the iterations

Here is my running code:

set.seed(123)
input.before.imp <- t(input)
data.imputed <- GS_impute(input.before.imp, n_cores = 96)

Thanks for your input..

Tai

[WandeRum]

Hello. I tested GSimp with my own dataset using parallel mode. Imputations went fine except I noticed the gibbs_res slot was empty:

str (data.imputed$gibbs_res)
 logi[1:3, 0 , 1:2000] 
 - attr(*, "dimnames")=List of 3
  ..$ : chr [1:3] "std" "yhat" "yres"
  ..$ : NULL
  ..$ : NULL

I wonder if there's anything wrong within the iterations

Here is my running code:

set.seed(123)
input.before.imp <- t(input)
data.imputed <- GS_impute(input.before.imp, n_cores = 96)

Thanks for your input..

Tai

Hi Tai. Yes, this sounds a bit weird. Can you compare your data with our example data on the repo. Also how large is your data set? Can you print dim(data)? It will be better to filter some features and samples with too many zeros before the imputation since it's really hard to infer any useful information from them.