Plot precision/recall/F1 versus confidence cutoff — plot_prec_recall_vs

Build a line plot of precision, recall, and F1 across confidence cutoffs from a PR-curve data frame (for example, output from compute_variant_pr_curve(), compute_tree_pr_curve(), or compute_pr_from_jaccard_mat()).

Usage

plot_prec_recall_vs_conf(
  pr_df,
  sample_name = NULL,
  cutoff = NULL,
  x_axis_scale = c("log10", "linear"),
  log10_min = 1e-06,
  j_thres = NULL,
  legend = TRUE
)

Arguments

pr_df: Data frame containing at least columns conf_cutoff, precision, recall, and f1. If present, fraction_well_matched is used for optional annotation.
sample_name: Optional title string shown above the panel.
cutoff: Optional numeric cutoff to highlight with a vertical dashed line and label near the top of the panel.
x_axis_scale: X-axis scale: "log10" (default) or "linear".
log10_min: Lower bound used for log10 plotting so zero cutoffs can be displayed safely.
j_thres: Optional Jaccard threshold used only for display annotation with fraction_well_matched.
legend: Logical; if TRUE, show metric legend on the right.

Value

A ggplot2 object.