Extract quantiles from a predictive CDF summary — extract

For each requested probability, finds the x value where the CDF (and its credible bounds) crosses that probability.

Usage

extract_quantiles(cdf_summary, probs = c(0.5, 0.95), cdf_mat = NULL)

Arguments

cdf_summary: A data frame produced by compute_predictive_cdf(), with columns x, median, low, and high.
probs: Numeric vector of probabilities to extract (default: c(0.5, 0.95)).
cdf_mat: Numeric matrix of posterior CDF draws as returned by compute_predictive_cdf() (rows = posterior draws, columns = x_seq grid points). When supplied, the 95% prediction interval bounds (x_low, x_high) are computed by interpolating each draw's CDF to find the x at which it crosses p, then taking the 2.5% and 97.5% quantiles across draws. This is fully consistent with the ribbon in the CDF plot (both derive from the same cdf_mat). If NULL, falls back to inverting the summary credible bands, which can fail near the tails.

Value

A data frame with columns quantile, quantile_label, x_low, and x_high. x_low and x_high are the 2.5% and 97.5% bounds of the 95% prediction interval for that quantile, on the same scale as the x column of cdf_summary.