binomial_glm_plot {model4you} | R Documentation |
Can be used on its own but is also useable as plotfun in
node_pmterminal
.
binomial_glm_plot(mod, data = NULL, plot_data = FALSE, theme = theme_classic(), ...)
mod |
A model of class glm with binomial family. |
data |
optional data frame. If NULL the data stored in mod is used. |
plot_data |
should the data in form of a mosaic type plot be plotted? |
theme |
A ggplot2 theme. |
... |
ignored at the moment. |
set.seed(2017) # number of observations n <- 1000 # balanced binary treatment # trt <- factor(rep(c("C", "A"), each = n/2), # levels = c("C", "A")) # unbalanced binary treatment trt <- factor(c(rep("C", n/4), rep("A", 3*n/4)), levels = c("C", "A")) # some continuous variables x1 <- rnorm(n) x2 <- rnorm(n) # linear predictor lp <- -0.5 + 0.5*I(trt == "A") + 1*I(trt == "A")*I(x1 > 0) # compute probability with inverse logit function invlogit <- function(x) 1/(1 + exp(-x)) pr <- invlogit(lp) # bernoulli response variable y <- rbinom(n, 1, pr) dat <- data.frame(y, trt, x1, x2) # logistic regression model mod <- glm(y ~ trt, data = dat, family = "binomial") binomial_glm_plot(mod, plot_data = TRUE) # logistic regression model tree ltr <- pmtree(mod) plot(ltr, terminal_panel = node_pmterminal(ltr, plotfun = binomial_glm_plot, confint = TRUE, plot_data = TRUE))