compute_calibration | R Documentation |
Estimate calibration given forecasts and corresponding outcomes
compute_calibration(
forecast,
outcome,
method = c("smoothing", "binning"),
CI = NULL,
binwidth = NULL,
...
)
forecast |
Vector of probability forecasts. |
outcome |
Vector of observations (0 or 1). |
method |
Method used to estimate calibration, either |
CI |
Confidence level (e.g. 0.95).
CI not computed if |
binwidth |
Binwidth when calibration is estimated by binning.
If |
... |
Arguments of |
Dataframe with columns Forecast
(bins), Frequency
(frequency of outcomes in the bin),
Lower
(lower bound of the CI) and Upper
(upper bound of the CI).
N <- 1e4
f <- rbeta(N, 1, 1)
o <- sapply(f, function(x) {rbinom(1, 1, x)})
lapply(c("binning", "smoothing"),
function(m) {
cal <- compute_calibration(f, o, method = m)
with(cal, plot(Forecast, Frequency, type = "l"))
abline(c(0, 1), col = "red")
})
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.