Credible Intervals and Point Estimates from Random Draws

Summarise the distribution of random draws in an rvec, using credible intervals and point estimates.

Usage

draws_ci(
  x,
  width = 0.95,
  prefix = NULL,
  na_rm = FALSE,
  point = c("median", "mean")
)

# S3 method for class 'rvec'
draws_ci(
  x,
  width = 0.95,
  prefix = NULL,
  na_rm = FALSE,
  point = c("median", "mean")
)

# S3 method for class 'rvec_chr'
draws_ci(
  x,
  width = 0.95,
  prefix = NULL,
  na_rm = FALSE,
  point = c("median", "mean")
)

Arguments

x

An rvec.

width

Width(s) of credible interval(s). One or more numbers greater than 0 and less than or equal to 1. Default is 0.95.

prefix

String to be added to the names of columns in the result. Defaults to name of x.

na_rm

Whether to remove NAs before calculating summaries. Default is FALSE.

point

Summary measure used for point estimate: "median" (the default) or "mean".

Value

A tibble with 2 * length(width) + 1 columns. When width has length 1, the columns are

• <name>.lower Lower limit of credible interval

• <name>.mid Point estimate

• <name>.upper Upper limit of the credible interval

Details

The credible intervals extend from the quantiles at 0.5 - 0.5 * width to the quantiles 0.5 + 0.5 * width.

When point is "mean", point estimates can lie outside the associated credible intervals, though this only happens in unusual cases, such as when distributions are highly skewed.

Warning

It is tempting to assign the results of a call to draws_ci() to a column in a data frame, as in

my_df$ci <- draws_ci(my_rvec)

However, creating columns in this way can corrupt an ordinary data frame. For safer options, see the examples below.

See also

draws_quantile() gives more options for forming quantiles.

Other ways of applying pre-specified functions across draws are:

• draws_all()

• draws_any

• draws_min()

• draws_max()

• draws_median()

• draws_mean()

• draws_mode()

• draws_sd()

• draws_var()

• draws_cv()

• draws_quantile()

Apply arbitrary function across draws:

• draws_fun()

Check for missing or finite values:

• draws_any_na()

• draws_all_na()

• draws_any_infinite()

• draws_all_infinite()

• draws_any_finite()

• draws_all_finite()

Examples

set.seed(0)
m <- rbind(a = rnorm(100, mean = 5, sd = 2),
           b = rnorm(100, mean = -3, sd = 3),
           c = rnorm(100, mean = 0, sd = 20))
x <- rvec(m)
x
#> <rvec_dbl<100>[3]>
#>                a                b                c 
#>     4.9 (2, 8.2) -3.6 (-7.1, 2.9)    1.1 (-35, 36) 
draws_ci(x)
#> # A tibble: 3 × 3
#>   x.lower x.mid x.upper
#>     <dbl> <dbl>   <dbl>
#> 1    2.02  4.93    8.23
#> 2   -7.06 -3.61    2.87
#> 3  -35.3   1.06   36.1 
draws_ci(x, width = c(0.5, 0.99))
#> # A tibble: 3 × 5
#>   x.lower x.lower1 x.mid x.upper1 x.upper
#>     <dbl>    <dbl> <dbl>    <dbl>   <dbl>
#> 1    1.21     3.86  4.93     6.25    9.85
#> 2   -7.81    -5.60 -3.61    -1.24    4.55
#> 3  -48.1    -12.3   1.06    13.0    47.4 
draws_ci(x, point = "mean")
#> # A tibble: 3 × 3
#>   x.lower x.mid x.upper
#>     <dbl> <dbl>   <dbl>
#> 1    2.02  5.05    8.23
#> 2   -7.06 -3.14    2.87
#> 3  -35.3   1.34   36.1 
draws_ci(x, prefix = "results")
#> # A tibble: 3 × 3
#>   results.lower results.mid results.upper
#>           <dbl>       <dbl>         <dbl>
#> 1          2.02        4.93          8.23
#> 2         -7.06       -3.61          2.87
#> 3        -35.3         1.06         36.1 

## results from 'draws_ci'
## assigned to a data frame
library(dplyr)
df <- data.frame(x)

## base R approach
cbind(df, draws_ci(x))
#>                  x    x.lower     x.mid   x.upper
#> 1     4.9 (2, 8.2)   2.017346  4.934077  8.228987
#> 2 -3.6 (-7.1, 2.9)  -7.058286 -3.611295  2.872534
#> 3    1.1 (-35, 36) -35.275211  1.061243 36.052985

## a tidyverse alternative:
## mutate with no '='
df |> mutate(draws_ci(x))
#>                  x    x.lower     x.mid   x.upper
#> 1     4.9 (2, 8.2)   2.017346  4.934077  8.228987
#> 2 -3.6 (-7.1, 2.9)  -7.058286 -3.611295  2.872534
#> 3    1.1 (-35, 36) -35.275211  1.061243 36.052985