initializeOperator: Initialise a proximal operator

In bozenne/lava.penalty: Penalized Latent Variable Models

Description

Generate the proximal operator for the requested penalty

Usage

initializeOperator(lambda1, index.penalty1, index.model1, lambda2,
  index.penalty2, index.model2, lambdaG, index.penaltyG, index.modelG, lambdaN,
  index.penaltyN, nrow, ncol, equivariance, constrain.variance, index.variance)

Arguments

lambda1	lasso penalization parameter.
index.penalty1	a list containing the position of lasso penalized parameters in the vector of parameters.
index.model1	the index of the model relative to each lasso penalized parameter
lambda2	ridge penalization parameter.
index.penalty2	a list containing the position of ridge penalized parameters in the vector of parameters.
index.model2	the index of the model relative to each ridge penalized parameter
lambdaG	group lasso penalization parameter.
index.penaltyG	a list containing the position of group lasso parameters in the vector of parameters.
index.modelG	the index of the model relative to each group lasso penalized parameter
lambdaN	nuclear norm penalization parameter.
index.penaltyN	a list containing the position of the parameters corresponding to each image in the vector of parameters.
nrow	a vector containing the number of rows of each image.
ncol	a vector containing the number of columnss of each image.
equivariance	should the lambda parameter be multiplied with the first variance parameter?
constrain.variance	should the variance parameters be exponential trans
index.variance	the position of the variance parameters in coef

Value

a list containing the proximal operator and the penalty funtion

Examples

lavaPenalty:::initializeOperator(lasso = TRUE, ridge = FALSE, groupLasso = FALSE, nuclearNorm = FALSE)

bozenne/lava.penalty documentation built on May 13, 2019, 1:41 a.m.