UMRgradDesc_fixed_df: Gradient Descent with a fixed number of constant pieces...

In UMR: Unmatched Monotone Regression

Description

Gradient Descent with a fixed number of constant pieces (degrees of freedom)

Usage

UMRgradDesc_fixed_df(
  grad,
  init,
  stepsize,
  MM,
  tol = 1e-07,
  printevery = Inf,
  filename
)

Arguments

grad	a function(mm) where mm is the previous iterate value (i.e., the estimate vector).
init	Initial value of estimate ('mm'). The output will be of length length(init).
stepsize	Gradient descent stepsize. Set carefully!
MM	Number of iterations in which "support reduction" (combining of approximately equal values into a region of constancy) is done (see details and paper). Depending on tol, may not use all MM iterations.
tol	Tolerance: end algorithm once sum(abs(mm-mmprev)) < tol or you hit MM iterations.
printevery	integer value (generally << MM). Every 'printevery' iterations, a count will be printed and the output saved.
filename	path1/path2/filename to save output to.

Details

UMRgradDesc_fixed_df does a gradient descent with a fixed (upper bound) on the number of constant segments of the function.

Output of UMRgradDesc_fixed_df is unsorted. Note weights for 'mm' are not passed in; rather they will be contained/used in grad().

UMR documentation built on Aug. 14, 2021, 9:09 a.m.