Calculate the entropy of a vector

Calculate the entropy of a vector

entropy(v, pseudo_count = length(v)/1000, na.action = na.fail)

np(v, p = 0.05, pseudo_count = p/5, na.action = na.fail)

modal_category(v, na.action = na.fail)

Arguments

v

categorical vector

pseudo_count

number of pseudo counts to add on, to stabilize empty categories

na.action

how to handle NA values

p

proportion threshold

Value

the sample entropy

Functions

• np: The number of categories exceeding p proportion of the total

• modal_category: The modal category of v. Ties are broken by lexicographic order of the factor levels.

Examples

v2 = gl(2, 4)
v4 = gl(4, 4)
stopifnot(entropy(v2) < entropy(v4))
v_empty = v2[1:4] #empty level 2
stopifnot(is.finite(entropy(v_empty))) # pseudo_count

np(v4, p = .2, pseudo_count = 0)
#> [1] 4
np(v4, p = .25, pseudo_count = 0)
#> [1] 0
np(v4, p = .25, pseudo_count = .0001)
#> [1] 4

modal_category(v4)
#> [1] "1"
modal_category(v4[-1])
#> [1] "2"