selectVariables: selectVariables

In valse: Variable Selection with Mixture of Models

Description

For a given lambda, construct the sets of relevant variables for each cluster.

Usage

selectVariables(
  phiInit,
  rhoInit,
  piInit,
  gamInit,
  mini,
  maxi,
  gamma,
  glambda,
  X,
  Y,
  thresh = 1e-08,
  eps,
  ncores = 3,
  fast
)

Arguments

phiInit	an initial estimator for phi (size: p*m*k)
rhoInit	an initial estimator for rho (size: m*m*k)
piInit	an initial estimator for pi (size : k)
gamInit	an initial estimator for gamma
mini	minimum number of iterations in EM algorithm
maxi	maximum number of iterations in EM algorithm
gamma	power in the penalty
glambda	grid of regularization parameters
X	matrix of regressors
Y	matrix of responses
thresh	real, threshold to say a variable is relevant, by default = 1e-8
eps	threshold to say that EM algorithm has converged
ncores	Number or cores for parallel execution (1 to disable)
fast	boolean to enable or not the C function call

Value

a list, varying lambda in a grid, with selected (the indices of variables that are selected), Rho (the covariance parameter, reparametrized), Pi (the proportion parameter)

valse documentation built on May 31, 2021, 9:10 a.m.