ADMM_EN_SMW: ADMM on l1 regularized quadratic program
In gumeo/accSDA: Accelerated Sparse Discriminant Analysis

ADMM_EN_SMW

R Documentation

ADMM on l1 regularized quadratic program

Applies Alternating Direction Method of Multipliers to the l1-regularized quadratic program

f(\mathbf{x}) + g(\mathbf{x}) = \frac{1}{2}\mathbf{x}^TA\mathbf{x} - d^T\mathbf{x} + \lambda |\mathbf{x}|_1

ADMM_EN_SMW(Ainv, V, R, d, x0, lam, mu, maxits, tol, quiet, selector)

`Ainv`	Diagonal of `A^{-1}` term in SMW formula, where A is an n by n positive definite coefficient matrix.
`V`	Matrix from SMW formula.
`R`	Upper triangular matrix in Cholesky decomposition of `I + UA^{-1}V`.
`d`	nx1 dimensional column vector.
`lam`	Regularization parameter for l1 penalty, must be greater than zero.
`mu`	Augmented Lagrangian penalty parameter, must be greater than zero.
`maxits`	Number of iterations to run
`tol`	Vector of stopping tolerances, first value is absolute, second is relative tolerance.
`quiet`	Logical controlling display of intermediate statistics.
`selector`	Vector to choose which parameters in the discriminant vector will be used to calculate the regularization terms. The size of the vector must be p the number of predictors. The default value is a vector of all ones. This is currently only used for ordinal classification.

This function is used by other functions and should only be called explicitly for debugging purposes.

ADMM_EN_SMW returns an object of class "ADMM_EN_SMW" including a list with the following named components