In Zadchow/concurve: Computes & Plots Compatibility (Confidence), Surprisal, & Likelihood Distributions

knitr::opts_chunk$set(echo = TRUE)

Suppose we wanted to produce confidence distributions for data with binary outcomes and where we employ a logistic regression, we would do the following. Here, I use the mtcars dataset for the example and also simulate some very simple binary data. We use suppressMessages() to avoid seeing the long list of profiling messages.

library(concurve)

X <- rnorm(100, mean = 0, sd = 1)
Y <- rbinom(n = 100, size = 1, prob = 0.5)

mydata1 <- data.frame(X, Y)


model1 <- glm(Y ~ X,
  data = mydata1,
  family = binomial(link = "logit"))

model2 <- glm(am ~ mpg, family = binomial(link = "logit"), data = mtcars)

summary(model1)
summary(model2)



model_pro <- suppressMessages(curve_gen(
  model = model1,
  var = "X",
  method = "glm",
  log = T,
  steps = 1000,
  table = TRUE))

model_con <- suppressMessages(curve_gen(
  model = model2,
  var = "mpg",
  method = "glm",
  log = T,
  steps = 1000,
  table = TRUE))

head(model_con[[1]], 10)


(ggcurve(model_con[[1]],
    measure = "ratio",
    type = "c",
    nullvalue = c(0.8, 1.2), title = "Confidence Curve",
    subtitle = "The function displays intervals at every level.",
    xaxis = expression(theta == ~"Range of Values"),
    yaxis1 = expression(paste(italic(p), "-value")),
    yaxis2 = "Levels for CI (%)"))


(ggcurve(model_pro[[1]],
    measure = "ratio",
    type = "c",
    nullvalue = c(0.8, 1.2), title = "Confidence Curve",
    subtitle = "The function displays intervals at every level.",
    xaxis = expression(theta == ~"Range of Values"),
    yaxis1 = expression(paste(italic(p), "-value")),
    yaxis2 = "Levels for CI (%)"))

(ggcurve(model_con[[2]],
    measure = "ratio",
    type = "cdf",
    nullvalue = c(0.8, 1.2), title = "Confidence Distribution",
    subtitle = "The function displays intervals at every level.",
    xaxis = expression(theta == ~"Range of Values"),
    yaxis1 = expression(paste(italic(p), "-value")),
    yaxis2 = "Levels for CI (%)"))

(ggcurve(model_pro[[2]],
    measure = "ratio",
    type = "cdf",
    nullvalue = c(0.8, 1.2), title = "Confidence Distribution",
    subtitle = "The function displays intervals at every level.",
    xaxis = expression(theta == ~"Range of Values"),
    yaxis1 = expression(paste(italic(p), "-value")),
    yaxis2 = "Levels for CI (%)"))


(ggcurve(model_con[[2]],
    measure = "ratio",
    type = "cd",
    nullvalue = NULL, title = "Confidence Density",
    subtitle = "The function displays intervals at every level.",
    xaxis = expression(theta == ~"Range of Values"),
    yaxis1 = expression(paste(italic(p), "-value")),
    yaxis2 = "Levels for CI (%)"))

(ggcurve(model_pro[[2]],
    measure = "ratio",
    type = "cd",
    nullvalue = NULL, title = "Confidence Density",
    subtitle = "The function displays intervals at every level.",
    xaxis = expression(theta == ~"Range of Values"),
    yaxis1 = expression(paste(italic(p), "-value")),
    yaxis2 = "Levels for CI (%)"))

Cite R Packages

Please remember to cite the R packages that you use in your work.

citation("concurve")
citation("cowplot")

---

References

---

Zadchow/concurve documentation built on Jan. 11, 2024, 4:55 a.m.