The Lasso Page
L1-constrained fitting
for statistics and data mining
|
|
The Lasso Page
L1-constrained fitting |
The Lasso is a shrinkage and selection method for linear regression.
It minimizes the usual sum of squared errors, with a bound on the sum of the
absolute values of the coefficients. It has connections to soft-thresholding
of wavelet coefficients, forward stagewise regression, and boosting methods.
The glmnet package for fitting Lasso and elastic net models can be found on
CRAN .
Here is a MATLAB version .
A simple explanation of the lasso and least angle regression
Lasso resources and links
Tibshirani, R. (1996). Regression shrinkage and selection via the lasso.
J. Royal. Statist. Soc B., Vol. 58, No. 1, pages 267-288).
Efron, B., Johnstone, I., Hastie, T. and Tibshirani, R. (2002). Least angle regression
pdf file.
Published in Annals of Statistics 2003
LARS software for Splus and R.
The software computes the entire LAR, Lasso or
Stagewise path in the same order of computations as a single least-squares fit.
Chen, Donoho, and Saunders: "Atomic Decomposition by Basis Pursuit(ps file)"
(pdf)
Elements of
Statistical Learning:
data mining, inference and prediction
(Hastie, Tibshirani ,Friedman)
Subset selection in regression (Miller)