cvFSGL: Fit and Cross-Validate a Linear Model with a Combination of...

In xhnnnnn/FSGL: Fit a Linear Model with a Combination of Lasso and Group Lasso Regularization. The calculation will more fast than 'SGL' package.

Description

Fit and Cross-Validate a Linear Model with a Combination of Lasso and Group Lasso Regularization

Usage

cvFSGL(
  data,
  index,
  nfold = 10,
  nlam = 20,
  min.frac = 0.05,
  alpha = 0.95,
  lambdas = NULL,
  thresh = 1e-04,
  maxit = 10000,
  gamma = 0.8,
  verbose = TRUE,
  step = 1,
  reset = 10,
  foldid = NA,
  UseUpperBound = FALSE
)

Arguments

data	data should be a list of $x$ and $y$, x is a data matrix (n x p) and y is a vector
index	A p-vector indicating group membership of each covariate
nfold	Number of folds of the cross-validation loop
nlam	Number of lambda to use in the regularization path
min.frac	The minimum value of the penalty parameter, as a fraction of the maximum value
alpha	The mixing parameter. alpha = 1 is the lasso penalty.
lambdas	A user inputted sequence of lambda values for fitting. We recommend leaving this NULL and letting FSGL self-select values
thresh	Convergence threshold for change in beta
maxit	Maximum number of iterations to convergence
gamma	Fitting parameter used for tuning backtracking (between 0 and 1)
verbose	Logical flag for whether or not step number will be output
step	Fitting parameter used for inital backtracking step size (between 0 and 1)
reset	Fitting parameter used for taking advantage of local strong convexity in nesterov momentum (number of iterations before momentum term is reset)
foldid	An optional user-pecified vector indicating the cross-validation fold in which each observation should be included. Values in this vector should range from 1 to nfold. If left unspecified, SGL will randomly assign observations to folds
UseUpperBound	A logical flag for using upper bound

Value

An object with S3 class "cv.FSGL"

lldiff

An nlam vector of cross validated negative log likelihoods (squared error loss in the linear case, along the regularization path)

llSD

An nlame vector of approximate standard deviations of lldiff

lambdas

The actual list of lambda values used in the regularization path.

fit

A model fit object created by a call to FSGL on the entire dataset

foldid

A vector indicating the cross-validation folds that each observation is assigned to

prevals

A matrix of prevalidated predictions for each observation, for each lambda-value

Examples

set.seed(1)
n = 50; p = 100; size.groups = 10
index <- ceiling(1:p / size.groups)
X = matrix(rnorm(n * p), ncol = p, nrow = n)
beta = (-2:2)
y = X[,1:5] %*% beta + 0.1*rnorm(n)
data = list(x = X, y = y)
cvFit = cvFSGL(data, index)

xhnnnnn/FSGL documentation built on Jan. 7, 2022, 6:51 p.m.