Balanced cross-validation segments

Internal function for constructing a class-balanced segment list for cross-validation (copy from a previous verson from package pamr).

balanced.folds(y, nfolds = min(min(table(y)), 10))

Arguments

y	A categorical vector of class memberships
nfolds	The required number of CV-segements or "folds".

Value

A list of sample segments.

Examples

data(BCdata)
balanced.folds(BCdata$Y)
#> [[1]]
#>  [1]  83 124  16  73 109  39  69 116  35   1 127  95
#> 
#> [[2]]
#>  [1]  12  94  60   9  22  56  21   4  41  84  18  93 103  86
#> 
#> [[3]]
#>  [1]  34 115 108  74  32  67  91  66  90 129  53  31 126
#> 
#> [[4]]
#>  [1]   2   5 119 113  46  81  45 112  47  48  72  75
#> 
#> [[5]]
#>  [1]  57  40  80 130  68  63  52 106  55  78  62  17 107
#> 
#> [[6]]
#>  [1] 105   3  54  29 100 120 118  27  19  50  96 125   8
#> 
#> [[7]]
#>  [1]  58  87 104  14  70  13  44  79 114   6  30  89  82
#> 
#> [[8]]
#>  [1]  76  37  23  99  43 121  33 128 101  36  97  71 111
#> 
#> [[9]]
#>  [1]  85  42  61 123 102  77   7 117  59  64  25  65  98  10
#> 
#> [[10]]
#>  [1]  20 122  49  88  28  92  26  38  11  51  24  15 110
#>