Ensemble of growth curves fitted to daily counts — fit_ensemble

Fits the Logistic, Gompertz and Richards incidence models by maximum likelihood and combines them with Akaike weights \(w_i \propto \exp(-\Delta\mathrm{AIC}_i/2)\) (weights = "akaike") or with inverse-MSE weights on the count scale (weights = "inverse_mse").

Usage

fit_ensemble_incidence(
  x,
  family = c("poisson", "negbin"),
  weights = c("akaike", "inverse_mse"),
  beta = 1.5,
  models = c("gompertz", "richards", "logistic"),
  warm = NULL
)

Arguments

x: Daily counts.
family: "poisson" or "negbin".
weights: Weighting scheme.
beta: Richards shape (held fixed).
models: Models included.
warm: Optional previous incidence_ensemble whose estimates are used as warm starts.

Value

An object of class incidence_ensemble.