Do manyROC analysis with cross-validation for hyperSpec object

[!!!] // No description yet //

sp_manyroc_with_cv(
  Var,
  Spectra,
  k_folds = 5,
  times = 10,
  seeds = 2222222,
  kind = "L'Ecuyer-CMRG",
  n_min = k_folds
)

Arguments

Var	(character(1)\ factor) Either the name of variable in Spectra which contains the grouping variable or a factor vector (or convertible to factor) with values for grouping.
Spectra	hyperSpec object, that contains colum ID for spectra blocking, columns with grouping variables as well as spectroscopic information.
k_folds	(positive integer) The number of folds in k-fold cross-validation.
times	(positive integer) The number of repetitions in repeated k-fold cross-validation.
seeds	(NULL | integer) Either a vector of integers of length times to set seed for each repetition of k-fold cross-validation. For more about seeds see set.seed(). Each seed will be passed to cvo_create_folds.()
kind	character or NULL. If kind is a character string, set R's RNG to the kind desired. Use "default" to return to the R default. See ‘Details’ for the interpretation of NULL.
n_min	integer(1) minimum acceptable number of unique samples per group. Must be at least k_folds or bigger.

Details

Function sp_manyroc_with_cv will be renamed in the future.

See also

sp_manyroc_with_cv_by_variable

Examples

library(manyROC)

fluorescence$ID  <- 1:nrow(fluorescence)
sp_manyroc_with_cv("gr", fluorescence[, , 500 ~ 502], k_folds = 3, times = 2)
#> $variable
#> [1] "gr"
#> 
#> $n_included
#> [1] 150
#> 
#> $ind_included_rows
#>  'as_str' logi [1:150] TRUE TRUE TRUE TRUE TRUE TRUE ...
#> 
#> $cvo
#> --- A cvo object: ----------------------------------------------------
#>  indices stratified blocked         cv_type k repetitions sample_size
#>    Train       TRUE   FALSE Repeated k-fold 3           2         150
#> 
#> Random number generator: L'Ecuyer-CMRG 
#> Seeds:  2222222, -30397663
#> ----------------------------------------------------------------------
#> 
#> $results
#>          set       fold compared_groups feature median_neg cutoff median_pos
#> 1   training Rep1_Fold1         A vs. B       1        341    335        324
#> 2   training Rep1_Fold1         A vs. C       1        341    306        344
#> 3   training Rep1_Fold1         B vs. C       1        324    343        344
#> 4   training Rep1_Fold2         A vs. B       1        338    348        334
#> ...      ...        ...             ...     ...        ...    ...        ...
#> 105     test Rep2_Fold2         B vs. C       3         NA    340         NA
#> 106     test Rep2_Fold3         A vs. B       3         NA    335         NA
#> 107     test Rep2_Fold3         A vs. C       3         NA    322         NA
#> 108     test Rep2_Fold3         B vs. C       3         NA    256         NA
#>                                                                             
#>      tp  fn  fp  tn sens spec  ppv  npv  bac youden kappa  auc
#> 1    26  11  14  21 0.70 0.60 0.65 0.66 0.65   0.30  0.30 0.63
#> 2    28   1  28   7 0.97 0.20 0.50 0.88 0.58   0.17  0.15 0.55
#> 3    16  13   8  29 0.55 0.78 0.67 0.69 0.67   0.34  0.34 0.66
#> 4    29   8  21  14 0.78 0.40 0.58 0.64 0.59   0.18  0.19 0.56
#> ... ... ... ... ...  ...  ...  ...  ...  ...    ...   ...  ...
#> 105   7   7   1  17 0.50 0.94 0.88 0.71 0.72   0.44  0.47   NA
#> 106  10   8   9   8 0.56 0.47 0.53 0.50 0.51   0.03  0.03   NA
#> 107   3  12   6  11 0.20 0.65 0.33 0.48 0.42  -0.15 -0.16   NA
#> 108   0  15   0  18 0.00 1.00  NaN 0.55 0.50   0.00  0.00   NA
#>                                                               
#>