Nowcast a reporting triangle into an ensemble (quasipoisson reporting regression)

A discriminative (regression) nowcast engine: for each horizon it regresses the settled total on the counts reported so far (`total ~ n[delay 0] + n[delay 1] + ...`, quasipoisson/identity, no intercept) and completes the incomplete weeks by simulating from that fit – parameter uncertainty plus a dispersion-matched negbin. No per-week magnitude parameter, so it is robust for the recent weeks and honestly dispersed. Shares the contract `f(reporting_triangle, ...) -> csfmt_ensemble_v3`.

Usage

nowcast_quasipoisson_v1(x, ...)

# S3 method for class 'csfmt_reporting_triangle_v3'
nowcast_quasipoisson_v1(
  x,
  max_delay,
  n_sim = 1000,
  denominator_col = NULL,
  delay_window = 26,
  ...
)

Arguments

x

A `csfmt_reporting_triangle_v3`.

...

Passed to methods.

max_delay

Delay horizon in weeks.

n_sim

Number of nowcast draws.

denominator_col

Optional denominator column to nowcast alongside.

delay_window

Train on only settled weeks within roughly this many weeks (tracks a drifting regime). Default 26; `NULL` uses all settled weeks.

Value

A `csfmt_ensemble_v3` with one row per reference week and an `n_sim`-column draw matrix of the nowcasted total per week (settled weeks degenerate at their observed total; incomplete weeks carry the regression's parameter + dispersion uncertainty). A second measure is added when `denominator_col` is given.