gettingtothebottom (version 3.0)

quantreg: Linear Programming - Quantile regression

Description

quantreg quantreg is used to fit quantile regression models.

Usage

quantreg(formula, tau = 0.5, lambda = 0)

Arguments

formula

An object of class "formula" (e.g., y~X)

tau

(optional) Quantile for quantile regression. Default is 0.5 but can also be a vector (e.g., tau = c(0.25,0.5,0.75)).

lambda

(optional) Regularization parameter. Default is lambda=0 (i.e., no regularization).

Examples

set.seed(12345)
n <- 20
p <- 20
X <- matrix(rnorm(n*p),n,p)
b0 <- double(p)
k <- 4
b0[sample(1:p,k,replace=FALSE)] <- 10*rnorm(k)
y <- X%*%b0 + 0.1*rnorm(n)

lambda <- 0
tau <- c(0.05,0.25,0.5,0.75,0.95)
sol <- quantreg(y~X,tau,lambda)
coef(sol)