hqreg-package {hqreg}

Regularization Paths for Lasso or Elastic-net Penalized Huber Loss Regression and Quantile Regression

Description

Efficient algorithms for fitting regularization paths for lasso or elastic-net penalized regression models with Huber loss, quantile loss or squared loss.

Details

Package:	hqreg
Type:	Package
Version:	1.4
Date:	2017-2-15
License:	GPL-3

Very simple to use. Accepts X,y data for regression models, and produces the regularization path over a grid of values for the tuning parameter lambda. Also provides functions for plotting, prediction and parallelized cross-validation.

Author(s)

Congrui Yi <[e-mail]>

References

Yi, C. and Huang, J. (2016) Semismooth Newton Coordinate Descent Algorithm for Elastic-Net Penalized Huber Loss Regression and Quantile Regression, https://arxiv.org/abs/1509.02957
Journal of Computational and Graphical Statistics, accepted in Nov 2016
http://www.tandfonline.com/doi/full/10.1080/10618600.2016.1256816

Examples

X = matrix(rnorm(1000*100), 1000, 100)
beta = rnorm(10)
eps = 4*rnorm(1000)
y = drop(X[,1:10] %*% beta + eps) 

# Huber loss
fit1 = hqreg(X, y)
coef(fit1, 0.01)
predict(fit1, X[1:5,], lambda = c(0.02, 0.01))
cv.fit1 = cv.hqreg(X, y)
plot(cv.fit1)

# Quantile loss
fit2 = hqreg(X, y, method = "quantile", tau = 0.2)
plot(fit2)

# Squared loss
fit3 = hqreg(X, y, method = "ls", preprocess = "rescale")
plot(fit3, xvar = "norm")