distributional_effects.Rd
Calculate distributional effects
distributional_effects(object, tails = "gaussian", ...) # S3 method for qs distributional_effects(object, tails = "gaussian", newdata = NULL, ...) # S3 method for matrix distributional_effects(object, tails, alphas, ...)
object | fit of class qs or matrix of fitted quantiles |
---|---|
tails | one of "gaussian" or "exponential" |
... | other parameters to pass |
newdata | new data to predict distributional outcomes for |
alphas | which quantiles these were fitted at |
The arguments alphas and quantiles are automatically handled if you pass an object of class "qs". If you don't pass new data to be predicted on, it will assume that you want to calculate distributional effects at the average of the data. This varies because of the non-linear model for the quantile process