Package edu.gsu.cs.dmlab.sparse.approximation

Class LARS_LassoSparseMatrixApproximator

java.lang.Object

edu.gsu.cs.dmlab.sparse.approximation.LARS_LassoCoeffVectorApproximator

edu.gsu.cs.dmlab.sparse.approximation.LARS_LassoSparseMatrixApproximator

All Implemented Interfaces:

ISparseMatrixApproximator, ISparseVectorApproximator

public class LARS_LassoSparseMatrixApproximator
extends LARS_LassoCoeffVectorApproximator
implements ISparseMatrixApproximator

Author:

Dustin Kempton, Data Mining Lab, Georgia State University

Constructor Summary

Constructors

Constructor	Description
LARS_LassoSparseMatrixApproximator(double lambda, LassoMode mode)	Constructor

Method Summary

Modifier and Type	Method	Description
smile.math.matrix.SparseMatrix	estimateCoeffs(smile.math.matrix.DenseMatrix signal, smile.math.matrix.DenseMatrix D)	Estimates the sparse coefficients of the input signal given the dictionary.

Methods inherited from class edu.gsu.cs.dmlab.sparse.approximation.LARS_LassoCoeffVectorApproximator

solve, solve

Methods inherited from class java.lang.Object

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

Constructor Detail

LARS_LassoSparseMatrixApproximator

public LARS_LassoSparseMatrixApproximator(double lambda,
                                          LassoMode mode)

Constructor

Parameters:

lambda - Depending on mode, lambda is used differently. If mode 1, then lambda min amount of correlation between a coefficient in A and signal b for the algorithm to proceed. If mode 2, max L1 norm of the coefficient vector.

mode - What mode to use PENALTY or L1COEFF. See lambda for differences.

Method Detail

estimateCoeffs

public smile.math.matrix.SparseMatrix estimateCoeffs(smile.math.matrix.DenseMatrix signal,
                                                     smile.math.matrix.DenseMatrix D)

Description copied from interface: ISparseMatrixApproximator

Estimates the sparse coefficients of the input signal given the dictionary.

Specified by:

estimateCoeffs in interface ISparseMatrixApproximator

Parameters:

signal - Matrix of input signals where each column is another m-dim signal

D - Dictionary of n-dim components used to reconstruct each m-dim signal

Returns:

Matrix of sparse coefficients used to reconstruct the signals using the dictionary