block.cd: A block coordinate descent function
In syhyunpark/csim: Constrained single index models (CSIM)

Description Usage Arguments Value See Also

block.cd implements the block-coordinate descent algorithm of Radchenko (2015) to optimize the (active) single index coefficients.

1
2
3

block.cd(w.a, ind.m, ind.old, u, y, Tr, X.a, d.link.fn.obj, it.max = 60,
  eps = 10^-4, trace = F, nbasis.t = c(6, 6, 8), rho.grid = c(0,
  0.25, 0.5), linear.link = FALSE, ortho.constr = TRUE)

`w.a`	a warm start (initial) value of the current nonzero single index coefficients.
`ind.m`	a value of the index used in defining each coordinate-wise block.
`ind.old`	a vector of the indices corresponding to the current nonzero coefficients.
`u`	a vector of the current singe index variable, u = alpha'X.
`y`	treatment outcomes, a n-by-1 vector.
`Tr`	treatment indicators, a n-by-1 vector; each element represents one of the K available treatment options.
`X.a`	a matrix of the pre-treatment covarates, associated with the current nonzero single index coefficients.
`d.link.fn.obj`	an object of the class `d.link.fn.obj` passed from the previous iteration.
`it.max`	an integer value specifying the maximum number of iterations for each coordinate.
`eps`	a value specifying the converge criterion of algorithm.
`trace`	if `TRUE`, show the trace of the fitting procedure; the default is `FALSE`.
`nbasis.t`	a length K+1 vector; each element specifies the number of B-spline basis funtions for approximating the treatment-specific link function; the last element is for the "main effect" link function; the default is `nbasis.t=NULL`, and will be determined depending on the sample size.
`rho.grid`	a grid vector of (ridge-type) smoothing parameters for approximating the link functions.
`linear.link`	if `TRUE`, restrict the link functions to be linear functions; the default is `FALSE`.
`ortho.constr`	the constraint that separates the interaction effects from the main effect (without this, the interaction effect can be confounded by the main effect); the default is `TRUE`.

a list of information of the optimized single index coefficients including

`w.a`	a vector of the optimized single index coefficients associated with the active index set.
`u`	a vector of the optimized single index variable.
`it`	the number of iterations.