Possible sets of modes

mode_possible_min() and mode_possible_max() determine the minimal and maximal sets of modes, given the number of missing values.

Usage

mode_possible_min(x, accept = FALSE, multiple = NULL)

mode_possible_max(x, accept = FALSE, multiple = NULL)

Arguments

x

A vector to search for its possible modes.

accept

Logical. If accept is set to TRUE, the functions don't necessarily return the minimal or maximal sets of modes but all values that might be part of those sets. Default is FALSE. See details.

multiple

Deprecated; will be removed in the future. Old name of accept.

Value

Vector of the same type as x. By default, it contains the minimal or maximal possible sets of modes (values tied for most frequent) in x. If the functions can't determine these possible modes because of missing values, they return NA.

Details

If accept = TRUE, the functions return multiple values that may or may not be modes depending on the true values behind NAs. Why is this disabled by default? In cases where multiple unique values would be modes if and only if one or more missing values represented them but there are not enough missing values to represent all of them, any values that are not represented by enough NAs would not be modes. This makes it unclear which unique values are part of the minimal and maximal sets of modes, so the default of accept is to return NA in these cases.

See also

mode_count_range() for the minimal and maximal numbers of possible modes. They can always be determined, even if the present functions return NA.

Examples

# "a" is guaranteed to be a mode,
# "b" might also be one, but
# "c" is impossible:
mode_possible_min(c("a", "a", "a", "b", "b", "c", NA))
#> [1] "a"
mode_possible_max(c("a", "a", "a", "b", "b", "c", NA))
#> [1] "a" "b"

# Only `8` can possibly be the mode
# because, even if `NA` is `7`, it's
# still less frequent than `8`:
mode_possible_min(c(7, 7, 8, 8, 8, 8, NA))
#> [1] 8
mode_possible_max(c(7, 7, 8, 8, 8, 8, NA))
#> [1] 8

# No clear minimal or maximal set
# of modes because `NA` may tip
# the balance between `1` and `2`
# towards a single mode:
mode_possible_min(c(1, 1, 2, 2, 3, 4, 5, NA))
#> [1] NA
mode_possible_max(c(1, 1, 2, 2, 3, 4, 5, NA))
#> [1] NA

# With `accept = TRUE`, the functions
# return all values that might be part of
# the min / max sets of modes; not these
# sets themselves:
mode_possible_min(c(1, 1, 2, 2, 3, 4, 5, NA), accept = TRUE)
#> [1] 1 2
mode_possible_max(c(1, 1, 2, 2, 3, 4, 5, NA), accept = TRUE)
#> [1] 1 2 3 4 5