Compute the confidence intervals of verification scores using a bootstrap method. If there is more than one forecast the confidence for the differences between forecasts is also computed.
Usage
bootstrap_score(
.fcst,
score_function,
parameter,
n,
groupings = "leadtime",
confidence_interval = 0.95,
...
)
Arguments
- .fcst
A
harp_fcst
object.- score_function
The name of the verification function to bootstrap for.
- parameter
The parameter that gives the name of the observations columns.
- n
Th number of bootstrap replicants.
- groupings
The groups to compute verification scores for. The default is leadtime.
- ...
Other arguments to the score function. e.g. thresholds.