getSignCorrection: workhorse for correctSigns

In deducorrect: Deductive Correction, Deductive Imputation, and Deterministic Correction

Description

Workhorse for correctSigns

Usage

getSignCorrection(r, A1, C1, eps, A2, C2, epsvec, flip, swap, w, swapIsOneFlip,
  maxActions, maxCombinations)

Arguments

r	The numerical record to correct
A1	Equality matrix
C1	Constant vector for equalities
eps	Tolerance for equality-checking
A2	Inequality matrix
C2	Constant vector for inequalities
epsvec	Vector to check against. (.Machine.double.eps for > inequalities, otherwise 0.)
flip	indices in r, where r may be sign-flipped
swap	n\times2 matrix with indices in r, each row indicating a possible value swap.
w	weight vector of length(flips)+nrow(swaps) if swapIsOneFlip==TRUE, otherwise of length(r)
swapIsOneFlip	logical. If TRUE, weights are assigned to each flip or swap action. If FALSE weights are assigned to every changed variable.
maxActions	Maximum number of flips+swaps to try.
maxCombinations	the maximum number of flip/swap combinations to try. See the description in correctSigns

Value

a list containing

S	n x length(r) array with corrected versions of r
weight	vector of length n with total weight for each solution
nFlip	number of sign flips for every solution
nSwap	number of value swaps for every solution

See Also

correctSigns

deducorrect documentation built on May 2, 2019, 3:47 p.m.