samoneval: Evaluates the loss function on a range of sigmas

In samon: Sensitivity Analysis for Missing Data

Evaluates the loss function on a range of sigmas

Description

Computes the loss function for a range of sigma

Usage

samoneval( mat, Npart = 10, sigmaList = c(1), type="both")

Arguments

mat	matrix with (i,j) entry representing value for subject i at time-point j.
Npart	Data is partitioned into Npart parts when estimating lossH and lossF.
sigmaList	vector of sigmas at which to evaluate the loss function.
type	one of "H", "F" or "both" to evaluate lossH, lossF or both for sigmaH or sigmaF in indicated range.

Details

Stepping through the values in sigmaList this functions computes the associated loss function. If type is set to "H" or "h", the loss function associated with dropout is computed. If type is set to "F" or "f" the loss function associated with outcome is computed. If type is set any other value then both types of loss function are evaluated.

Value

Returns an N by 2 matrix of results if type is "H" or "F" or an N by 4 matrix if type is "both". There is one row for each value of sigma. If type is "H" then the columns are sigmaH and lossH, if type is "F" then the columns are sigmaF and lossF, and, if type = "both" then the columns are sigmaH, lossH, sigmaF, and lossF.

Examples

data("samonPANSS1")
## dropout loss function
H1 <- samoneval( mat = samonPANSS1, Npart = 5,
      sigmaList      = seq(1,10,by=1),          
      type           = "H" )

samon documentation built on Oct. 26, 2023, 9:06 a.m.