isodistrreg (version 0.1.0)

qscore: Quantile scores for IDR or raw forecasts

Description

Computes quantile scores of IDR quantile predictions or of quantile predictions from raw forecasts in a data.frame.

Usage

qscore(predictions, quantiles, y)

Arguments

predictions

either an object of class idr (output of predict.idrfit), or a data.frame of numeric variables. In the latter case, quantiles are computed using the empirical distribution of the variables in predictions.

quantiles

numeric vector of desired quantiles.

y

a numeric vector of obervations of the same length as the number of predictions, or of length 1. In the latter case, y will be used for all predictions.

Value

A matrix of the quantile scores for the desired quantiles, one column per quantile.

Details

The quantile score of a forecast x for the u-quantile is defined as $$ 2(1{x > y} - u)(x - y), $$ where y is the observation. For u = 1/2, this equals the mean absolute error of the median forecast.

References

Gneiting, T. and Raftery, A. E. (2007), 'Strictly proper scoring rules, prediction, and estimation', Journal of the American Statistical Association 102(477), 359-378

See Also

predict.idrfit, qpred

Examples

# NOT RUN {
data("rain")

## Postprocess HRES forecast using data of 3 years

X <- rain[1:(3 * 365), "HRES", drop = FALSE]
y <- rain[1:(3 * 365), "obs"]

fit <- idr(y = y, X = X)

## Compute mean absolute error of the median postprocessed forecast using
## data of the next 2 years (out-of-sample predictions) and compare to raw
## HRES forecast

data <- rain[(3 * 365 + 1):(5 * 365), "HRES", drop = FALSE]
obs <- rain[(3 * 365 + 1):(5 * 365), "obs"]

predictions <- predict(fit, data = data)
idrMAE <- mean(qscore(predictions, 0.5, obs))
rawMAE <- mean(qscore(data, 0.5, obs))

c("idr" = idrMAE, "raw" = rawMAE)
# }