Construct Consonance Functions from Any CI-Producing Function

Source: R/curve_wrap.R

A generic wrapper that constructs consonance, surprisal, and likelihood functions from any function that produces confidence intervals. This is the most flexible approach for generating consonance curves from arbitrary statistical procedures.

curve_wrap(ci_func, steps = 1000, cores = getOption("mc.cores", 1L),
  table = TRUE)

Arguments

ci_func

A function that takes a confidence level (0-1) as its first argument and returns a numeric vector of length 2 (lower, upper bounds). See examples for common patterns.

steps

Number of consonance levels to compute. Default is 1000.

cores

Number of cores for parallel computation. Default uses getOption("mc.cores", 1L).

table

Logical. If TRUE (default), includes a summary table of key intervals in the output.

Value

A list with class "concurve" containing:

Intervals Dataframe

Data frame with columns: lower.limit, upper.limit, intrvl.width, intrvl.level, cdf, pvalue, svalue

Intervals Density

Data frame for plotting density functions

Intervals Table

Summary table of key intervals (if table = TRUE)

Details

This function repeatedly calls ci_func at different confidence levels to construct the full consonance function. The ci_func argument must be a function that:

  1. Takes a confidence level (numeric between 0 and 1) as its first argument

  2. Returns a numeric vector of exactly length 2: c(lower_bound, upper_bound)

Common patterns for ci_func:

  • Linear models: function(level) confint.default(model, parm = "x", level = level)

  • GLMs (profile): function(level) confint(model, parm = "x", level = level)

  • t-tests: function(level) t.test(x, conf.level = level)$conf.int

  • Correlations: function(level) cor.test(x, y, conf.level = level)$conf.int

  • Proportions: function(level) prop.test(x, n, conf.level = level)$conf.int

  • Survival: function(level) exp(confint(coxph_model, level = level)["var",])

See also

ggcurve() for plotting

curve_gen() for model-specific consonance functions

curve_rev() for constructing curves from published intervals

curve_table() for interval summaries

Examples

if (FALSE) { # \dontrun{
# Example 1: Linear model
data(mtcars)
model <- lm(mpg ~ wt + hp, data = mtcars)
ci_func <- function(level) confint.default(model, parm = "wt", level = level)
result <- curve_wrap(ci_func)
ggcurve(result[[1]], type = "c")

# Example 2: t-test
x <- rnorm(30, mean = 5)
ci_func <- function(level) t.test(x, conf.level = level)$conf.int
result <- curve_wrap(ci_func)
ggcurve(result[[1]], type = "c", nullvalue = 0)

# Example 3: Correlation
x <- rnorm(50)
y <- x + rnorm(50)
ci_func <- function(level) cor.test(x, y, conf.level = level)$conf.int
result <- curve_wrap(ci_func)

# Example 4: GLM with profile likelihood CIs
model <- glm(am ~ wt, data = mtcars, family = binomial)
ci_func <- function(level) confint(model, parm = "wt", level = level)
result <- curve_wrap(ci_func)

# Example 5: Proportion test
ci_func <- function(level) prop.test(45, 100, conf.level = level)$conf.int
result <- curve_wrap(ci_func)

# Example 6: Custom function with additional arguments
my_ci <- function(level, data, method) {
  cor.test(data$x, data$y, method = method, conf.level = level)$conf.int
}
df <- data.frame(x = rnorm(50), y = rnorm(50))
ci_func <- function(level) my_ci(level, data = df, method = "spearman")
result <- curve_wrap(ci_func)
} # }