set_parameters: Set parameters for tsmvr_solve

In spcorum/tsmvr: Truly Sparse Multivariate Regression

Description

Sets paramters for tsmvr_solve and functions that call tsmvr_solve (tsmvr_cv, tsmvr_replicate, and tsmvr_gridsearch).

Usage

set_parameters(B_type = c("gd", "ls"), Omega_type = c("gd", "ls",
  "min"), eta1 = 0.1, eta2 = 0.2, lam1 = 0.1, lam2 = 0.1,
  del1 = 1e-06, del2 = 1e-06, del3 = 1e-06, rho1 = 1e-06,
  rho2 = 1e-06, beta1 = 0.5, beta2 = 0.5, qmax1 = 128,
  qmax2 = 128, eps1 = 1e-04, eps2 = Inf, k = 5, reps = 10,
  max_iter = 2000, skip = 10, quiet = FALSE, suppress = FALSE,
  disp_min_ev = FALSE, save_history = FALSE)

Arguments

B_type	type of descent for regression steps (string: 'gd' or 'ls')
Omega_type	(string: 'gd', 'ls' or 'min')
eta1	B-step learning rate (positive numeric)
eta2	Omega-step learning rate (positive numeric)
lam1	B-step learning rate (non-negative numeric)
lam2	Omega-step learning rate (non-negative numeric)
del1	B initialization matrix inversion fudge factor (non-negative numeric)
del2	Omega initialization matrix inversion fudge factor (non-negative numeric)
del3	Omega step matrix inversion fudge factor (non-negative)
rho1	B-step linesearch convergence parameter (positive numeric)
rho2	Omega-step linesearch convergence parameter (positive numeric)
beta1	B-step linesearch stepsize shrinking paramter (positive numeric)
beta2	Omega-step linesearch stepsize shrinking paramter (positive numeric)
qmax1	B-step linesearch maximum number of iterations (positive integer valued numeric)
qmax2	Omega-step linesearch maximum number of iterations (positive integer valued numeric)
eps1	B-step convergence parameter (positive numeric or infinity)
eps2	Omega-step convergence parameter (positive numeric or infinity)
k	number of cross-validation folds (positive integer valued numeric greater than 1)
reps	number of k-fold cross-validation replicates (positive integer valued numeric)
max_iter	maximum number of iterations (positive integer)
skip	iteration skip frequency for output to screen (positive integer)
quiet	whether or not to operate (bool)
suppress	whether or not to suppress warnings (bool)
disp_min_ev	display the minimum value of the Omega iterate for each iteration (bool)
save_history	whethor or not to save and return the histories of the B and Omega iterates (bool)

Value

Returns a named list of input parameters.

See also tsmvr

References

\insertRef

chen2016hightsmvr \insertRefchen2018covariatetsmvr

spcorum/tsmvr documentation built on Aug. 31, 2019, 8:58 p.m.