is.control {islasso}

Auxiliary for controlling islasso model fitting

Description

Auxiliary function for controlling the islasso model fitting.

Usage

is.control(sigma2 = -1, tol = 1E-05, itmax = 1E+3, stand = TRUE,
  trace = 0, nfolds = 5, seed = NULL, adaptive = FALSE, g = .5,
  b0 = NULL, V0 = NULL, c = .5)

Arguments

sigma2

optional. The fixed value of dispersion parameter. If -1 (default) it is estimated from the data

tol

tollerance value to declare convergence, dafault to 1e-5

itmax

maximum number of iterations, default to 1000

stand

if TRUE (default), the covariates are standardized prior to fitting the model. However the coefficients are always returned on the original scale.

trace

Should the iterative procedure be printed? 0: no printing, 1 = compact printing, 2 = enlarged printing, 3 = compact printing including Fisher scoring information (only used in glm family).

nfolds

if lambda is unspecified in islasso, the number of folds to be used to perform cross valdation. Default to 5, and nfolds>2 is allowed. Although nfolds can be as large as the sample size (leave-one-out CV), it is not recommended for large datasets. nfolds is ignored if lambda is supplied.

seed

optional, the seed to be used to split the dataframe and to perform cross validation. Useful to make reproducible the results.

adaptive

experimental, if TRUE the adaptive LASSO is implemented.

g

a value belonging to the interval [0, 1]. Classical BIC is returned by letting g = 0 (default value), whereas extended BIC corresponds to the case g = 0.5.

b0

optional, starting values for the regression coefficients. If NULL, the point estimates from glmnet are used.

V0

optional, starting value for the estimates covariance matrix, If NULL, the identity matrix is used.

c

the weight of the mixture in the induced smoothed lasso, the default is c = .5. c = -1 means to compute it at each step of the iterative algorithm.

Author(s)

Maintainer: Gianluca Sottile <[e-mail]>