R recipes check_missing 检查缺失值

check_missing 创建配方操作的规范，该规范将检查变量是否包含缺失值。

用法

check_missing(
  recipe,
  ...,
  role = NA,
  trained = FALSE,
  columns = NULL,
  skip = FALSE,
  id = rand_id("missing")
)

参数

recipe: 一个菜谱对象。该检查将添加到该配方的操作序列中。
...: 一个或多个选择器函数用于选择用于此检查的变量。有关更多详细信息，请参阅selections()。
role: 由于没有创建新变量，因此此检查未使用。
trained: ... 中的选择器是否已由 prep() 解析的逻辑。
columns: 所选变量名称的字符串。该字段是一个占位符，一旦使用 prep() 就会被填充。
skip: 一个合乎逻辑的。当bake() 烘焙食谱时是否应该跳过检查？虽然所有操作都是在 prep() 运行时烘焙的，但某些操作可能无法对新数据进行(例如处理结果变量)。使用skip = TRUE时应小心，因为它可能会影响后续操作的计算。
id: 此检查唯一的字符串，用于识别它。

值

recipe 的更新版本，将新检查添加到任何现有操作的序列中。

细节

如果任何检查列确实包含 NA 值，则此检查将破坏 bake 函数。如果检查通过，则数据不会发生任何更改。

tidy() 结果

当您tidy()进行此检查时，将返回带有terms列(所选的选择器或变量)的tibble。

也可以看看

其他检查：check_class()、check_cols()、check_new_values()、check_range()

例子

data(credit_data, package = "modeldata")
is.na(credit_data) %>% colSums()
#>    Status Seniority      Home      Time       Age   Marital   Records 
#>         0         0         6         0         0         1         0 
#>       Job  Expenses    Income    Assets      Debt    Amount     Price 
#>         2         0       381        47        18         0         0 

# If the test passes, `new_data` is returned unaltered
recipe(credit_data) %>%
  check_missing(Age, Expenses) %>%
  prep() %>%
  bake(credit_data)
#> # A tibble: 4,454 × 14
#>    Status Seniority Home     Time   Age Marital Records Job       Expenses
#>    <fct>      <int> <fct>   <int> <int> <fct>   <fct>   <fct>        <int>
#>  1 good           9 rent       60    30 married no      freelance       73
#>  2 good          17 rent       60    58 widow   no      fixed           48
#>  3 bad           10 owner      36    46 married yes     freelance       90
#>  4 good           0 rent       60    24 single  no      fixed           63
#>  5 good           0 rent       36    26 single  no      fixed           46
#>  6 good           1 owner      60    36 married no      fixed           75
#>  7 good          29 owner      60    44 married no      fixed           75
#>  8 good           9 parents    12    27 single  no      fixed           35
#>  9 good           0 owner      60    32 married no      freelance       90
#> 10 bad            0 parents    48    41 married no      partime         90
#> # ℹ 4,444 more rows
#> # ℹ 5 more variables: Income <int>, Assets <int>, Debt <int>,
#> #   Amount <int>, Price <int>

# If your training set doesn't pass, prep() will stop with an error
if (FALSE) {
recipe(credit_data) %>%
  check_missing(Income) %>%
  prep()
}

# If `new_data` contain missing values, the check will stop `bake()`

train_data <- credit_data %>% dplyr::filter(Income > 150)
test_data <- credit_data %>% dplyr::filter(Income <= 150 | is.na(Income))

rp <- recipe(train_data) %>%
  check_missing(Income) %>%
  prep()

bake(rp, train_data)
#> # A tibble: 1,338 × 14
#>    Status Seniority Home   Time   Age Marital Records Job       Expenses
#>    <fct>      <int> <fct> <int> <int> <fct>   <fct>   <fct>        <int>
#>  1 bad           10 owner    36    46 married yes     freelance       90
#>  2 good           0 rent     60    24 single  no      fixed           63
#>  3 good           1 owner    60    36 married no      fixed           75
#>  4 good           8 owner    60    30 married no      fixed           75
#>  5 good          19 priv     36    37 married no      fixed           75
#>  6 good          15 priv     24    52 single  no      freelance       35
#>  7 good          33 rent     24    68 married no      freelance       65
#>  8 good           5 owner    60    22 single  no      fixed           45
#>  9 good          19 owner    60    43 single  no      fixed           75
#> 10 good          15 owner    36    43 married no      fixed           75
#> # ℹ 1,328 more rows
#> # ℹ 5 more variables: Income <int>, Assets <int>, Debt <int>,
#> #   Amount <int>, Price <int>
if (FALSE) {
bake(rp, test_data)
}

源代码：R/missing.R

相关用法

注：本文由纯净天空筛选整理自Max Kuhn等大神的英文原创作品 Check for Missing Values。非经特殊声明，原始代码版权归原作者所有，本译文未经允许或授权，请勿转载或复制。