Poisson model for incidence data

Usage

poisson_gam_model.incidence(
  d = i_incidence_input,
  model_fn = gam_poisson_model_fn(...),
  ...,
  frequency = "1 day",
  predict = list(),
  ip = i_discrete_ip,
  quick = FALSE,
  .progress = interactive()
)

Arguments

model_fn

a function that takes data relating to one time series (e.g. the input data d on a group by group basis) and returns a fitted GAM. The default creates a simple poisson model based on count alone (gam_poisson_model_fn()).

...

Named arguments passed on to gam_poisson_model_fn

window: controls the knot spacing in the GAM (if the default)
knots_fn: a function that takes the data as an input and returns a set of integers as time points for GAM knots, for s(time) term. The default here provides a roughly equally spaced grid determined by window, by a user supplied function could do anything. The input this function is the raw dataframe of data that will be considered for one model fit. It is guaranteed to have at least a time and count column. It is possible to

predict

if the GAM model in model_fn introduces other variables we need to know what their values should be fixed at for prediction. This is a named list of defaults for variables in the model supplied by model_fn. These defaults will be used in prediction. This may be supplied as part of the model function generator ( e.g. gam_delayed_reporting(...)$predict). If this is set to exactly FALSE no prediction is performed and a list column of fitted GAM models returned instead.