layer_quantile_distn: Returns predictive quantiles

In cmu-delphi/epipredict: Basic epidemiology forecasting methods

Returns predictive quantiles

Description

This function calculates quantiles when the prediction was distributional.

Usage

layer_quantile_distn(
  frosting,
  ...,
  quantile_levels = c(0.05, 0.1, 0.25, 0.5, 0.75, 0.9, 0.95),
  truncate = c(-Inf, Inf),
  name = ".pred_distn",
  id = rand_id("quantile_distn")
)

Arguments

frosting	a frosting postprocessor
...	Unused, include for consistency with other layers.
quantile_levels	a vector of probabilities to extract
truncate	Do we truncate the distribution to an interval
name	character. The name for the output column.
id	a random id string

Details

Currently, the only distributional modes/engines are

• quantile_reg()

• smooth_quantile_reg()

• rand_forest(mode = "regression") %>% set_engine("grf_quantiles")

If these engines were used, then this layer will grab out estimated (or extrapolated) quantiles at the requested quantile values.

Value

an updated frosting postprocessor. An additional column of predictive quantiles will be added to the predictions.

Examples

jhu <- covid_case_death_rates %>%
  filter(time_value > "2021-11-01", geo_value %in% c("ak", "ca", "ny"))

r <- epi_recipe(jhu) %>%
  step_epi_lag(death_rate, lag = c(0, 7, 14)) %>%
  step_epi_ahead(death_rate, ahead = 7) %>%
  step_epi_naomit()

wf <- epi_workflow(r, quantile_reg(quantile_levels = c(.25, .5, .75))) %>%
  fit(jhu)

f <- frosting() %>%
  layer_predict() %>%
  layer_quantile_distn() %>%
  layer_naomit(.pred)
wf1 <- wf %>% add_frosting(f)

p <- forecast(wf1)
p

cmu-delphi/epipredict documentation built on March 5, 2025, 12:17 p.m.