R recipes step_unknown 將缺失的類別分配給“未知”

step_unknown(
  recipe,
  ...,
  role = NA,
  trained = FALSE,
  new_level = "unknown",
  objects = NULL,
  skip = FALSE,
  id = rand_id("unknown")
)

dummy_names()

其他虛擬變量和編碼步驟：step_bin2factor() , step_count() , step_date() , step_dummy_extract() , step_dummy_multi_choice() , step_dummy() , step_factor2string() , step_holiday() , step_indicate_na() , step_integer() , step_novel() , step_num2factor() , step_ordinalscore() , step_other() , step_regex() , step_relevel() , step_string2factor() , step_time() , step_unorder()

data(Sacramento, package = "modeldata")

rec <-
  recipe(~ city + zip, data = Sacramento) %>%
  step_unknown(city, new_level = "unknown city") %>%
  step_unknown(zip, new_level = "unknown zip") %>%
  prep()

table(bake(rec, new_data = NULL) %>% pull(city),
  Sacramento %>% pull(city),
  useNA = "always"
) %>%
  as.data.frame() %>%
  dplyr::filter(Freq > 0)
#>               Var1            Var2 Freq
#> 1         ANTELOPE        ANTELOPE   33
#> 2           AUBURN          AUBURN    5
#> 3     CAMERON_PARK    CAMERON_PARK    9
#> 4       CARMICHAEL      CARMICHAEL   20
#> 5   CITRUS_HEIGHTS  CITRUS_HEIGHTS   35
#> 6             COOL            COOL    1
#> 7  DIAMOND_SPRINGS DIAMOND_SPRINGS    1
#> 8        EL_DORADO       EL_DORADO    2
#> 9  EL_DORADO_HILLS EL_DORADO_HILLS   23
#> 10       ELK_GROVE       ELK_GROVE  114
#> 11         ELVERTA         ELVERTA    4
#> 12       FAIR_OAKS       FAIR_OAKS    9
#> 13          FOLSOM          FOLSOM   17
#> 14      FORESTHILL      FORESTHILL    1
#> 15            GALT            GALT   21
#> 16   GARDEN_VALLEY   GARDEN_VALLEY    1
#> 17      GOLD_RIVER      GOLD_RIVER    4
#> 18     GRANITE_BAY     GRANITE_BAY    3
#> 19       GREENWOOD       GREENWOOD    1
#> 20         LINCOLN         LINCOLN   22
#> 21          LOOMIS          LOOMIS    2
#> 22          MATHER          MATHER    1
#> 23    MEADOW_VISTA    MEADOW_VISTA    1
#> 24 NORTH_HIGHLANDS NORTH_HIGHLANDS   21
#> 25      ORANGEVALE      ORANGEVALE   11
#> 26          PENRYN          PENRYN    1
#> 27     PLACERVILLE     PLACERVILLE   10
#> 28   POLLOCK_PINES   POLLOCK_PINES    3
#> 29  RANCHO_CORDOVA  RANCHO_CORDOVA   28
#> 30  RANCHO_MURIETA  RANCHO_MURIETA    3
#> 31       RIO_LINDA       RIO_LINDA   13
#> 32         ROCKLIN         ROCKLIN   17
#> 33       ROSEVILLE       ROSEVILLE   48
#> 34      SACRAMENTO      SACRAMENTO  438
#> 35    WALNUT_GROVE    WALNUT_GROVE    1
#> 36 WEST_SACRAMENTO WEST_SACRAMENTO    3
#> 37          WILTON          WILTON    5

tidy(rec, number = 1)
#> # A tibble: 1 × 3
#>   terms value        id           
#>   <chr> <chr>        <chr>        
#> 1 city  unknown city unknown_NRDyG