estimateAUCwithPairwiseCML {BLOQ}

estimate AUCwith pairwise censored maximum likelihood

Description

function to estimate mean and and covariance matrix of censored data using a full censored maximum likelihood approach via fitting all possible pairs, then use these estimates for estimating AUC and its standard error

Usage

estimateAUCwithPairwiseCML(
  inputData,
  LOQ,
  timePoints,
  isMultiplicative = FALSE,
  onlyFitCML = FALSE,
  optimizationMethod = NULL,
  CMLcontrol = NULL,
  na.rm = TRUE
)

Arguments

inputData

numeric matrix or data frame of the size n by J (n the sample size and J the number of time points) the input dataset

LOQ

scalar, limit of quantification value

timePoints

vector of time points

isMultiplicative

logical variable indicating whether an additive error model (FALSE) or a multiplicative error model (TRUE) should be used

onlyFitCML

logical variable with FALSE as default, if TRUE only the censored maximum likelihood estimates will be calculated.

optimizationMethod

single string specifying the method to be used for optimizing the log-likelihood, the default is NULL that allows the function to decide the about the best method. Otherwise, one can select among choices available via R package maxLik: "NR" (for Newton-Raphson), "BFGS" (for Broyden-Fletcher-Goldfarb-Shanno), "BFGSR" (for the BFGS algorithm implemented in R), "BHHH" (for Berndt-Hall-Hall-Hausman), "SANN" (for Simulated ANNealing), "CG" (for Conjugate Gradients), or "NM" (for Nelder-Mead). Lower-case letters (such as "nr" for Newton-Raphson) are allowed.

CMLcontrol

list of arguments to control convergence of maximization algorithm. It is the same argument as control in the function maxLik in the R package maxLik

na.rm

logical variable indicating whether the lines with missing values should be ignored (TRUE, default) or not (FALSE). Note that, it will be applied for the sub-datasets regarding each pair.

Value

a list with three components: output of maxLik function, estimated parameters (mean vector and the covariance matrix) using censored maximum likelihood, and estimated AUC and its standard error.

Author(s)

Vahid Nassiri, Helen Yvette Barnett

See Also

maxLik

Examples

# generate data from Beal model with only fixed effects
set.seed(111)
genDataFixedEffects <- simulateBealModelFixedEffects(10, 0.693,
		1, 1, seq(0.5,3,1.5))
estimateAUCwithPairwiseCML(genDataFixedEffects, 0.1, seq(0.5,3,1.5))