Skip to contents

Distribution function of an interval-valued variable

Source: R/interval.R
tda_idf.Rd

tda_idf: the distribution and density function of a continuous interval-valued variable, as bounds at each partition point – an interval variable does not pin down a single distribution, only a range one is consistent with. tda_iddf is the same idea for a discrete interval-valued variable (a finite set of possible values, not a continuum), and, unlike idf, does take opt= for a self-consistent distribution; idf takes only fmt=.

Usage

tda_idf(
  formula,
  data,
  self_consistent = FALSE,
  control = NULL,
  options = list(),
  dir = tempfile("tda"),
  ...
)

tda_iddf(
  formula,
  data,
  self_consistent = FALSE,
  control = NULL,
  options = list(),
  dir = tempfile("tda"),
  ...
)

Arguments

formula

a one-sided formula naming one interval-valued variable, ~ iv(lo, hi).

data

a data frame.

self_consistent

also compute the self-consistent distribution (opt=2), not just the min/max/mean bounds (opt=1, default) – opt= of both commands. For tda_idf it comes back as $self_consistent (partition point and distribution function) with $converged; the fixed point satisfies the self-consistency equation to the tolerance (checked in the tests against the equation iterated in R).

control

convergence settings (mxit=/tolf via tda_control), defaults 50 and 0.001.

options

a named list of further TDA options, passed through.

dir

working directory.

...

passed to tda_run.

Value

An object carrying a table of partition points and the distribution function's bounds and mean at each.

See also

Other interval-valued data: tda_bounds(), tda_ilsreg(), tda_imean(), tda_imreg(), tda_inpreg(), tda_ivar1(), tda_ivreg(), tda_sddf()

Examples

set.seed(1)
n <- 30
d <- data.frame(lo = round(rnorm(n, 5, 2), 1))
d$hi <- d$lo + round(runif(n, 0.5, 2), 1)
fit <- tda_idf(~ iv(lo, hi), d)
head(fit$table)
#>   index partition      lower      upper    mean_df
#> 1     1       0.6 0.00000000 0.03333333 0.00000000
#> 2     2       1.0 0.00000000 0.06666667 0.01333333
#> 3     3       1.6 0.03333333 0.06666667 0.05333333
#> 4     4       2.0 0.06666667 0.06666667 0.06666667
#> 5     5       2.1 0.06666667 0.10000000 0.06666667
#> 6     6       2.8 0.10000000 0.10000000 0.10000000

# iddf: the discrete counterpart, with the self-consistent option idf
# itself does not have
fit2 <- tda_iddf(~ iv(lo, hi), d, self_consistent = TRUE)
head(fit2$table)
#>   index      lower      upper    mean_df self_consistent_df
#> 1     0 0.00000000 0.03333333 0.01666667       3.483206e-05
#> 2     1 0.00000000 0.06666667 0.03333333       4.994775e-02
#> 3     2 0.03333333 0.06666667 0.05000000       5.001742e-02
#> 4     3 0.00000000 0.13333333 0.06111111       1.392193e-05
#> 5     4 0.03333333 0.30000000 0.14444444       2.290678e-01
#> 6     5 0.00000000 0.36666667 0.15000000       3.318755e-03