Summarize distribution with center

Source: R/summ_center.R

Functions to compute center of distribution. summ_center() is a wrapper for respective summ_*() functions (from this page) with default arguments.

summ_center(f, method = "mean")

summ_mean(f)

summ_median(f)

summ_mode(f, method = "global")

summ_midrange(f)

Arguments

f

A pdqr-function representing distribution.
method

Method of center computation. For summ_center() is one of "mean", "median", "mode", "midrange". For summ_mode() is one of "global" or "local".

Value

summ_center(), summ_mean(), summ_median() and summ_mode(*, method = "global") always return a single number representing a center of distribution. summ_mode(*, method = "local") can return a numeric vector with multiple values representing local maxima.

Details

summ_mean() computes distribution's mean.

summ_median() computes a smallest x value for which cumulative probability is not less than 0.5. Essentially, it is a as_q(f)(0.5). This also means that for pdqr-functions with type "discrete" it always returns an entry of "x" column from f's "x_tbl" metadata.

summ_mode(*, method = "global") computes a smallest x (which is an entry of "x" column from f's x_tbl) with the highest probability/density. summ_mode(*, method = "local") computes all x values which represent non-strict local maxima of probability mass/density function.

summ_midrange() computes middle point of f's support (average of left and right edges).

See also

summ_spread() for computing distribution's spread, summ_moment() for general moments.

Other summary functions: summ_classmetric(), summ_distance(), summ_entropy(), summ_hdr(), summ_interval(), summ_moment(), summ_order(), summ_prob_true(), summ_pval(), summ_quantile(), summ_roc(), summ_separation(), summ_spread()

Examples

# Type "continuous"
d_norm <- as_d(dnorm)
## The same as `summ_center(d_norm, method = "mean")`
summ_mean(d_norm)

#> [1] -1.358047e-16
summ_median(d_norm)

#> [1] 2.292681e-14
summ_mode(d_norm)

#> [1] -2.904343e-12
## As pdqr-functions always have finite support, output here is finite
summ_midrange(d_norm)

#> [1] -2.904788e-12

# Type "discrete"
d_pois <- as_d(dpois, lambda = 10)
summ_mean(d_pois)

#> [1] 9.999985
summ_median(d_pois)

#> [1] 10
## Returns the smallest `x` with highest probability
summ_mode(d_pois)

#> [1] 9
## Returns all values which are non-strict local maxima
summ_mode(d_pois, method = "local")

#> [1]  9 10
## As pdqr-functions always have finite support, output here is finite
summ_midrange(d_pois)

#> [1] 14

# Details of computing local modes
my_d <- new_d(data.frame(x = 11:15, y = c(0, 1, 0, 2, 0) / 3), "continuous")
## Several values, which are entries of `x_tbl`, are returned as local modes
summ_mode(my_d, method = "local")

#> [1] 12 14