sprm (version 1.1)

sprm-package: Sparse and Non-Sparse Partial Robust M Regression

Description

In this package two methods for dimension reduction and regression analysis are implemented that yield estimates with a partial least squares alike interpretability. Partial robust M regression is robust to both vertical outliers and leverage points. Sparse partial robust M regression is a related robust method with sparse coefficient estimate, and therefore with intrinsic variable selection.

Details

ll{ Package: sprm Type: Package Version: 1.1 Date: 2014-12-10 License: GPL(>=3) } The main functions in this package are prms and sprms for non-spares and sparse partial robust M regression, respectively. Further cross validation procedures for tuning parameter selection are implemented in prmsCV and sprmsCV. Various plot options are available to visualize the results.

References

Sven Serneels et al. (2014) Sparse partial robust M regression

Serneels, S., Croux, C., Filzmoser, P., Van Espen, P.J., Partial Robust M-Regression. Chemometrics and Intelligent Laboratory Systems, 79 (2005), 55-64.

See Also

prms, sprms

Examples

set.seed(50235)
U1 <- c(rep(3,20), rep(4,30))
U2 <- rep(3.5,50)
X1 <- replicate(5, U1+rnorm(50))
X2 <- replicate(20, U2+rnorm(50))
X <- cbind(X1,X2)
beta <- c(rep(1, 5), rep(0,20))
e <- c(rnorm(45,0,1.5),rnorm(5,-20,1))
y <- X%*%beta + e
d <- as.data.frame(X)
d$y <- y
mod <- prms(y~., data=d, a=2, fun="Hampel")
smod <- sprms(y~., data=d, a=2, eta=0.5, fun="Hampel")

biplot(mod)
biplot(smod)