A specific parameter or set of parameters can be estimated by a pipeline.
This function applies the pipeline to a synthetic epidemic with sawtooth
incidence resulting from a stepped growth rate function. The lag between
synthetic input and estimate is assessed by minimising the root mean square
error of input and estimated based on different lag offsets.
Arguments
- pipeline
a function taking an input dataset and an infectivity profile as inputs and producing an estimate as output. This is the whole parametrised pipeline including any other inputs. This can be a
purrrstyle function, in which case the.xvariable is the input dataset and.yis the infectivity profile.- ip
the infectivity profile. - a dataframe with columns:
boot (anything + default(1)) - a bootstrap identifier
probability (proportion) - the probability of new event during this period.
a0 (double) - the beginning of the time period (in days)
a1 (double) - the end of the time period (in days)
Minimally grouped by: boot (and other groupings allowed).
A default value is defined.
- lags
a vector with the delays to test. Defaults to -10 to +30 days
Value
a lag analysis dataframe containing the estimate type and the lag
in days that the estimate is behind the actual observation
Examples
# lags from a locfit incidence model with Rt estimation.
# This model has no estimator lag:
pipeline = ~ .x %>% poisson_locfit_model() %>% rt_from_incidence(ip = .y)
quantify_lag(pipeline, ip = test_ip)
#> Rt from incidence: assuming independence and approximating quantiles.
#> (N.B. this message will only be displayed once.)
#> # A tibble: 3 × 2
#> # Groups: estimate [3]
#> estimate lag
#> * <chr> <int>
#> 1 growth 0
#> 2 incidence 0
#> 3 rt 0
# lags from an epiestim Rt estimation
# this model's lags depend on the infectivity profile.
# In this case it is 8 days
pipeline2 = ~ .x %>% rt_epiestim(ip = .y)
quantify_lag(pipeline2, ip=test_ip )
#> # A tibble: 1 × 2
#> # Groups: estimate [1]
#> estimate lag
#> * <chr> <int>
#> 1 rt 8