Root Confidence Intervals for Every Datestamped Episode

Convenience wrapper that runs explosive_root/root_ci on every episode in a datestamp result, so root inference doesn't have to be hand-run per episode.

Usage

root_ci_datestamp(object, ds, level = 0.95, type = c("normal", "cauchy"))

Arguments

object

A radf_obj (or subclass) that ds was computed on – needs the original data, retrieved via its "mat" attribute.

ds

A datestamp result computed on object. Set min_duration there to exclude episodes too short for reliable root inference.

level

Confidence level, passed to root_ci (default 0.95).

type

CI type, passed to root_ci (default "normal").

Value

A named list (one element per series in ds; the panel sieve-bootstrap case, whose ds entry is named "panel" and has no single corresponding series, is dropped with a warning), each a data frame with one row per datestamped episode: Start, End, rho, rho_lower, rho_upper, doubling_time, doubling_time_lower, doubling_time_upper.

Details

Not folded into summary.radf_obj: that function's existing S3 dispatch (summary_radf.mc_cv/.wb_cv/.sb_cv) is built entirely around radf_cv test-statistic critical values – root CIs are a different kind of output (needing a datestamp() result, not a radf_cv) with no natural fit in that dispatch chain. A standalone function keeps this addition to the size the enhancement notes actually scoped ("a small follow-up"), rather than restructuring summary()'s shared machinery to accommodate a fundamentally different kind of result.

Root inference on a very short episode is statistically meaningless (the same way it would be calling explosive_root directly on 2-3 points) – this function doesn't filter episodes itself, since datestamp already has a min_duration argument for exactly this; set it there before piping into this function, rather than expecting this function to second-guess what counts as "too short".

Status

See also

explosive_root, root_ci, datestamp