Skip to content

Calculate Binomial Confidence Interval

calc_ci_binom.Rd

Calculates the joint binomial confidence interval based on the binomial variance given the data. Uses normal approximation of the binomial.

Usage

calc_ci_binom(p, n, ci = sqrt(0.95))

Arguments

p

numeric(1). The classification metric in [0, 1]. Can also be a vector of values representing the metric of interest (sens or spec).

n

integer(1). The total number of counts in the denominator for the metric being calculated.

ci

numeric(1). The width of the confidence interval to be calculated. Must be in [0.5, 1].

Value

A tibble object of the upper and lower binomial confidence limits corresponding to the value of ci.

References

The Statistical Evaluation of Medical Tests for Classification and Prediction. 2004. Margaret Pepe, Altman, DG, Bland, JM. 1994. "Diagnostic tests 1: sensitivity and specificity", British Medical Journal, vol 308, 1552. (I think?).

See also

qnorm()

Author

Stu Field

Examples

tp <- 16
fn <- 4
sens <- tp / (tp + fn)
calc_ci_binom(sens, tp + fn)
#> # A tibble: 1 × 2
#>   lower upper
#>   <dbl> <dbl>
#> 1 0.600     1