optimismBoot error

[Weekend-Warrior]

I'm getting an error in the nominalTrainWorkflow. .export in the foreach loop of result is looking for an object called 'optimismBoot', assigned just before the loop but no object exists. Is there a missing condition?

[topepo]

Please read the issue template. Providing a reproducible example and the results of sessionInfo will help get your question answered.

[Weekend-Warrior]

Sure Max, thanks. I apologize for the breach of etiquette.

[topepo]

I'm putting together a CRAN submission, so if you have an example to send, sooner is better than later.

[alanocallaghan]

I have run into this problem also. I believe it is related to the use of parallel backends.

library(caret)
library(doParallel)
library(glmnet)

set.seed(100)

cl <- makeCluster(2)
registerDoParallel(cl)

predictor_matrix <- matrix(rnorm(10000), ncol=200, 
    dimnames=list(NULL, paste("column", 1:200)))
labels <- factor(sample(1:0, nrow(predictor_matrix), replace=TRUE))

tctl <- trainControl(
    method = "repeatedcv",
    number = 5,
    repeats = 10,
    allowParallel = TRUE
)

fit1 <- caret::train(
    x = predictor_matrix,
    y = labels,
    method = "glmnet",
    family = "binomial",
    trControl = tctl
)

Output:

Error in e$fun(obj, substitute(ex), parent.frame(), e$data) : 
  unable to find variable "optimismBoot"

sessionInfo()
R version 3.4.1 (2017-06-30)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Ubuntu 16.04.3 LTS

Matrix products: default
BLAS: /usr/lib/openblas-base/libblas.so.3
LAPACK: /usr/lib/libopenblasp-r0.2.18.so

locale:
 [1] LC_CTYPE=en_GB.UTF-8       LC_NUMERIC=C               LC_TIME=en_GB.UTF-8        LC_COLLATE=en_GB.UTF-8    
 [5] LC_MONETARY=en_GB.UTF-8    LC_MESSAGES=en_GB.UTF-8    LC_PAPER=en_GB.UTF-8       LC_NAME=C                 
 [9] LC_ADDRESS=C               LC_TELEPHONE=C             LC_MEASUREMENT=en_GB.UTF-8 LC_IDENTIFICATION=C       

attached base packages:
[1] parallel  stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
[1] glmnet_2.0-10      Matrix_1.2-11      doParallel_1.0.10  iterators_1.0.8    foreach_1.4.3      caret_6.0-77      
[7] ggplot2_2.2.1.9000 lattice_0.20-35    devtools_1.13.3   

loaded via a namespace (and not attached):
 [1] Rcpp_0.12.12       ddalpha_1.2.1      gower_0.1.2        bindr_0.1          compiler_3.4.1     DEoptimR_1.0-8    
 [7] plyr_1.8.4         class_7.3-14       tools_3.4.1        rpart_4.1-11       ipred_0.9-6        digest_0.6.12     
[13] lubridate_1.6.0    memoise_1.1.0      tibble_1.3.3       nlme_3.1-131       gtable_0.2.0       pkgconfig_2.0.1   
[19] rlang_0.1.1        RcppRoll_0.2.2     prodlim_1.6.1      bindrcpp_0.2       e1071_1.6-8        withr_2.0.0       
[25] stringr_1.2.0      dplyr_0.7.2        recipes_0.1.0      stats4_3.4.1       nnet_7.3-12        CVST_0.2-1        
[31] grid_3.4.1         glue_1.1.1         robustbase_0.92-7  R6_2.2.2           survival_2.41-3    lava_1.5          
[37] purrr_0.2.3        reshape2_1.4.2     kernlab_0.9-25     magrittr_1.5       DRR_0.0.2          splines_3.4.1     
[43] scales_0.4.1.9002  codetools_0.2-15   ModelMetrics_1.1.0 MASS_7.3-47        assertthat_0.2.0   dimRed_0.1.0      
[49] timeDate_3012.100  colorspace_1.3-2   stringi_1.1.5      lazyeval_0.2.0     munsell_0.4.3    

[topepo]

I can reproduce it. I'll look into it.

[topepo]

That should fix the issue. Please retest and check on your system.

[alanocallaghan]

Works fine for me now, thanks for the quick fix!

[bvvamsik]

I am still having this issue and still getting the error
Error in e$fun(obj, substitute(ex), parent.frame(), e$data) :
unable to find variable "optimismBoot"

[YinanZheng]

Same problem here as @bvvamsik . Using latest 6.0.77. I also still can reproduce the problem using the @alanocallaghan posted above.

[alanocallaghan]

Are you guys running the github version or the CRAN version? The latest CRAN version was pushed before this was fixed.

[Triamus]

Installing from GitHub via devtools::install_github('topepo/caret/pkg/caret') solved it for me.

[bvvamsik]

"GitHub Installation" Worked perfectly for me too, hope we will get that CRAN repo update as well, which might help us while scaling our machine learning algorithms..

[garycai]

After installing caret from github, I still ran into this problem when using the recipes packages:

library(caret)
library(doParallel)
library(glmnet)
library(recipes)

set.seed(100)

cl <- makeCluster(2)
registerDoParallel(cl)

test_data <- as.data.frame(matrix(rnorm(10000), ncol=200, 
                                  dimnames=list(NULL, paste("column", 1:200))))
test_data$label <- factor(sample(1:0, nrow(test_data), replace=TRUE))

data_recipe <- recipe(label ~ ., data = test_data) %>%
    step_center(all_predictors()) %>%
    step_scale(all_predictors())

tctl <- trainControl(
    method = "repeatedcv",
    number = 5,
    repeats = 10,
    allowParallel = TRUE
)

fit1 <- caret::train(
    x = data_recipe,
    data = test_data,
    method = "glmnet",
    family = "binomial",
    trControl = tctl
)

Output
Error in e$fun(obj, substitute(ex), parent.frame(), e$data) : unable to find variable "optimism_boot"

[tomehta]

Any idea when the fix will be committed to CRAN ? I am getting the same error while trying parallel processing with caret

[YinanZheng]

@tomehta @garycai I am using R 3.4.2 and installing from github devtools::install_github('topepo/caret/pkg/caret') works pretty well for me... I had an impression that CRAN update will be around December.

[topepo]

I'm shooting for this week.

[caa25]

Ran into exact same problem today (11/19/17). Triggered devtools::install_github('topepo/caret/pkg/caret'). Works like a charm. Thanks all.

[topepo]

@tomehta @alanocallaghan @bvvamsik

FYI... A new version has been sent to CRAN but a problem with a reverse dependency is holding it up. I've contacted the package owner but have not heard back. I'll give it a few more days then ask them to let it through.

[mbeekink]

@topepo Thank you for solving the issue. Although the improved version is still nog available through CRAN (28/11/2017). When do you think it will be available?