powered by
The model approach use scaled lasoo approach for model selection.
robregcc_option( maxiter = 10000, tol = 1e-10, nlam = 100, out.tol = 1e-08, lminfac = 1e-08, lmaxfac = 10, mu = 1, nu = 1.05, sp = 0.3, gamma = 2, outMiter = 3000, inMiter = 500, kmaxS = 500, tolS = 1e-04, nlamx = 20, nlamy = 20, spb = 0.3, spy = 0.3, lminfacX = 1e-06, lminfacY = 0.01, kfold = 10, fullpath = 0, sigmafac = 2 )
maximum number of iteration for convergence
tolerance value set for convergence
number of lambda to be genrated to obtain solution path
tolernce value set for convergence of outer loop
a multiplier of determing lambda_min as a fraction of lambda_max
a multiplier of lambda_max
penalty parameter used in enforcing orthogonality
penalty parameter used in enforcing orthogonality (incremental rate of mu)
maximum proportion of nonzero elements in shift parameter
adaptive penalty weight exponential factor
maximum number of outer loop iteration
maximum number of inner loop iteration
maximum number of iteration for fast S estimator for convergence
tolerance value set for convergence in case of fast S estimator
number of x lambda
number of y lambda
sparsity in beta
sparsity in shift gamma
a multiplier of determing lambda_min as a fraction of lambda_max for sparsity in X
a multiplier of determing lambda_min as a fraction of lambda_max for sparsity in shift parameter
nummber of folds for crossvalidation
1/0 to get full path yes/no
multiplying factor for the range of standard deviation
a list of controling parameter.
# NOT RUN { # default options library(robregcc) control_default = robregcc_option() # manual options control_manual <- robregcc_option(maxiter=1000,tol = 1e-4,lminfac = 1e-7) # }
Run the code above in your browser using DataLab