Create a Log-Odds Plot — plot_log

Plot the log-odds, $$log(Prob / (1 - Prob))$$ for each sample. See Section for note about extreme probabilities.

Usage

plot_log_odds(
  truth,
  predicted,
  pos_class,
  cutoff = 0.5,
  y_lab = NULL,
  max_prob = NULL,
  scramble = FALSE
)

Arguments

truth: character(n) or factor(n). A vector of true class names. In most instances you will have to also pass a pos_class argument defining the positive/event class.
predicted: numeric(n). A numeric vector of class probabilities.
pos_class: character(1). Name of the "positive" or "event" class.
cutoff: numeric(1). A cutoff for the decision/operating point, predictions above which are considered the positive class.
y_lab: character(1). Optional label for the y-axis.
max_prob: numeric(1). Experimental. Maximum probability value cutoff for the log-odds plot. Removes extreme samples from the plot to avoid distorting x-axis.
scramble: logical(1). Should values be randomized to avoid monotonically decreasing probability scores (aesthetic)?

Extreme probabilities

extreme values in [0, 1] are thresholded at .Machine$double.eps^0.5, or [1.490116119e-08, 0.9999999851] to restrict the x-axis and avoid Inf values in log-odds space (0/1).

Author

Stu Field

Examples

n <- 20
withr::with_seed(22, {
  true <- sample(c("control", "disease"), n, replace = TRUE)
  pred <- runif(n)
})
plot_log_odds(true, pred, "disease")

plot_log_odds(true, pred, "disease", scramble = TRUE)