[R] quantile regression and moments

[jacinthe at gmx.de]

Hi,

how to derive an estimate of skewness and kurtosis out of a predicted distribution by quantile regression?

Example:
library(quantreg)
data(airquality)
airq <- airquality[143,]
f <- rq(Ozone ~ ., data=airquality,tau=seq(0.01,0.99,0.01))
predict(f,newdata=airq)

Any suggestions?

Kind regards,
Jaci
--