mpredcc: Fit Mean and Predictor Covariance-Connected (MPREDCC)...
In koekvall/mpredcc: Fully Likelihood-based Principal Components Regression

Description Usage Arguments Value

Fit Mean and Predictor Covariance-Connected (MPREDCC) regression model

mpredcc(Y, X, Z = NULL, H = NULL, k = 1, mu = 1e-06, tol = 1e-06,
  maxit = 1000, quiet = TRUE, relative = TRUE, maxit_upd = c(50,
  50), quiet_upd = c(TRUE, TRUE), tol_upd = c(1e-06, 1e-06),
  accelerate = TRUE, restart = TRUE, step_size = NULL, Psi = NULL,
  alpha = NULL, ssx = NULL, ssy = NULL, theta = NULL, phi = NULL)

`Y`	An n-vector of responses
`X`	An n x p matrix of predictors.
`Z`	Optional n x p_y design matrix for responses
`H`	Optional np x p_x design matrix for predictors
`k`	The dimension of the dimension reduction subspace
`mu`	Ridge penalty coefficient for regression coefficient
`tol`	Tolerance for termianting the algorithm
`maxit`	Maximum number of iterations in coordinate descent
`quiet`	If false, print progress after every coordinate update
`relative`	If true, use relative change to determine convergence
`maxit_upd`	Maximum iterations for Psi and ssx updates (length = 2)
`quiet_upd`	If false, print progress within Psi and ssx updates (length = 2)
`tol_upd`	Tolerance for Psi and ssx updates (length = 2)
`accelerate`	If true, will use acceleration in Psi update, see apg::apg
`restart`	If true, use adaptive restart in Psi update, see apg::apg
`step_size`	Initial step_size for Psi update, see apg::apg
`Psi`	Starting value for Psi (ignored if is.null(ssx))
`alpha`	Starting value for alpha
`ssx`	Starting value for ssx (ignored if is.null(Psi))
`ssy`	Starting value for ssy
`theta`	Starting value for theta (ignored if is.null(Z))
`phi`	starting value for phi (ignored if is.null(H))