dev.nm.stat.regression.linear.lasso

Class ConstrainedLASSOProblem

java.lang.Object

dev.nm.stat.regression.linear.LMProblem

dev.nm.stat.regression.linear.lasso.ConstrainedLASSOProblem

public class ConstrainedLASSOProblem
extends LMProblem

A LASSO (least absolute shrinkage and selection operator) problem focuses on solving an RSS (residual sum of squared errors) problem with L1 regularization. The constrained form solves \[ \min_w (\left \| Xw - y \right \|_2^2), \left \| w \right \|_1 \leq t \]

See Also:

• R. Tibshirani, "Regression shrinkage and selection via the LASSO," Journal of the Royal Statistical Society, Series B, Volume: 58, Issue: 1, 267 - 288, 1996.
• Wikipedia: LASSO method

Constructor Summary

Constructors

Constructor and Description
ConstrainedLASSOProblem(ConstrainedLASSOProblem that) Copy constructor.
ConstrainedLASSOProblem(Vector y, Matrix X, double t) Constructs a LASSO problem in the constrained form.

Method Summary

Modifier and Type	Method and Description
double	t() Get the penalization parameter for the constrained form of LASSO.

Methods inherited from class dev.nm.stat.regression.linear.LMProblem

A, checkInputs, intercept, invOfwAtwA, nExogenousFactors, nFactors, nObs, wA, weights, wy, X, y

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

ConstrainedLASSOProblem

public ConstrainedLASSOProblem(Vector y,
                               Matrix X,
                               double t)

Constructs a LASSO problem in the constrained form.

Parameters:

y - the vector of response variable (n * 1), properly demeaned and scaled

X - the design matrix of factors (n * m), properly demeaned and scaled

t - the penalization parameter

ConstrainedLASSOProblem

public ConstrainedLASSOProblem(ConstrainedLASSOProblem that)

Copy constructor.

Parameters:

that - another LASSOProblem

Method Detail

t

public double t()

Get the penalization parameter for the constrained form of LASSO.

Returns:

t, the penalization parameter for the constrained form of LASSO