ggplot.train level plots has incorrectly reordered axis

[MarekOzana]

It seems that ggplot.train is converting the tuneGrid parameters into factors and thus sometimes creating cases when numerical order is broken. Consequently the charts have incorrect axis order.
e.g. instead of 0.0001, 0.01, 1 it produces 1e-2, 1e-4, 1 (string order not number order).

Minimal, reproducible example:

library(caret)
data(mtcars)
m <- train(mpg ~ cyl + disp,
           data = mtcars,
           method="svmRadial",
           tuneGrid = expand.grid(C=1:2, sigma=c(0.0001, 0.01, 1)))
ggplot(m, plotType="level")

the produced chart has incorrectly ordered sigma axes.

Session Info:

>sessionInfo()
R version 3.3.2 (2016-10-31)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows >= 8 x64 (build 9200)

locale:
[1] LC_COLLATE=English_United States.1252  LC_CTYPE=English_United States.1252   
[3] LC_MONETARY=English_United States.1252 LC_NUMERIC=C                          
[5] LC_TIME=English_United States.1252    

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

other attached packages:
[1] kernlab_0.9-25  caret_6.0-73    ggplot2_2.2.1   lattice_0.20-34

loaded via a namespace (and not attached):
 [1] Rcpp_0.12.10       magrittr_1.5       splines_3.3.2      MASS_7.3-45        munsell_0.4.3     
 [6] colorspace_1.3-2   foreach_1.4.3      minqa_1.2.4        stringr_1.2.0      car_2.1-4         
[11] plyr_1.8.4         tools_3.3.2        parallel_3.3.2     nnet_7.3-12        pbkrtest_0.4-7    
[16] grid_3.3.2         gtable_0.2.0       nlme_3.1-131       mgcv_1.8-17        quantreg_5.29     
[21] MatrixModels_0.4-1 iterators_1.0.8    digest_0.6.12      lme4_1.1-12        lazyeval_0.2.0    
[26] assertthat_0.1     tibble_1.2         Matrix_1.2-8       nloptr_1.0.4       reshape2_1.4.2    
[31] ModelMetrics_1.1.0 codetools_0.2-15   labeling_0.3       stringi_1.1.3      compiler_3.3.2    
[36] scales_0.4.1       stats4_3.3.2       SparseM_1.76      

[m-dz]

I think I have a fix ready, just need to figure out how to submit it...

In the meantime, one more example:

x <- train(mpg ~ cyl + disp,
           data = mtcars,
           method="xgbTree",
           tuneGrid = expand.grid(
             nrounds=c(20),
             max_depth=c(2),
             eta = c(0.0001, 0.01, 1),
             gamma = c(0.0001, 0.01, 1),
             colsample_bytree=c(0.0001, 0.01, 1),
             min_child_weight = 1,
             subsample = 1
           ))

ggplot(x, plotType="level", nameInStrip = TRUE)

[m-dz]

I have added a pull request #624 as an attempt to fix this issue. Both @MarekOzana's and my examples produce expected plots. I hope I did not break anything...