clustDRM (version 0.1-0)

simulEvalDRM: simulation-based evalutation of a dose-response model

Description

a function to simulate data based different dose-response model using parameters estimated from a provided pilot study. The function then simulate data from the estimated model for the given dose levels and number of replications per dose. Some criteria will be compauted which then can be used to compare different settings.

Usage

simulEvalDRM(pilotData, doseLevels, numReplications, numSim,
  standardDeviation, EDp = 0.5, funcList = c("linear", "linlog",
  "exponential", "emax", "sigEmax", "logistic", "betaMod", "quadratic"))

Arguments

pilotData

a dataset presenting dose-response data from a pilot study. The first column should give the doses and the second one should give the response values.

doseLevels

the dose levels which should be used in the simulation study.

numReplications

number of replications for each of thes dose levels for the simulated data.

numSim

number of times that the simulation study should be replicated.

standardDeviation

standard deviation of the generated response.

EDp

scalar in (0,1), indicatign with EDp should be computed to compare different models, default is 0.5 (ED50).

funcList

string vector with models for data generation and fitting, should be selected from c("linear", "linlog", "exponential", "emax", "sigEmax", "logistic", "betaMod","quadratic").

Value

a list with the following elements

estEDp a list of length of funcList providing the estimated EDp from models fitted to data generated from each model in funcList realEDp a vector of length funcList, the EDp's computed based on the estimated parameters from different models fitted to pilotData bestModel a list of length funcList, a frequency table of best selected model for data generated from each model in funcList meanEDp a matrix showing mean of estimated EDp's averaged over numSim replications. biasEDp a matrix showing bias of estimated EDp's averaged over numSim replications. mseEDp a matrix showing MSE of estimated EDp's averaged over numSim replications. varEDp a matrix showing variance of estimated EDp's averaged over numSim replications. relativeBiasEDp a matrix showing relative bias of estimated EDp's averaged over numSim replications. absBiasEDp a matrix showing absolute bias of estimated EDp's averaged over numSim replications. absRelativeBiasEDp a matrix showing absolute bias of estimated EDp's averaged over numSim replications. averagaedAIC a matrix showing AIC's of different models averaged over numSim replications. quantity2Plot which if needed will be passed to plot method.

The output of simulEvalDRM can be passed to the function plotSimulDRM to plot a heatmap for the desired the quantity of interest. Possible quantities are ("mean", "bias", "mse", "variance", "relativeBias", "absBias", "absRelativeBias")

Examples

# NOT RUN {
## gnerating data, a sample of size 20
set.seed(11)
doses2Use <-  c(0, 5, 20)
numRep2Use <- c(3, 3, 3)
generatedData <- cbind(rep(1,sum(numRep2Use)),
MCPMod::genDFdata("logistic",c(5, 3, 10, 0.05), doses2Use,
 numRep2Use, 1), 
		matrix(rnorm(1*sum(numRep2Use)), sum(numRep2Use), 1))
colnames(generatedData) <- c("ID", "dose", "response", "x1")
for (iGen in 2:20){
genData0 <- cbind(rep(iGen,sum(numRep2Use)), 
MCPMod::genDFdata("logistic",c(5, 3, 10, 0.05), doses2Use, 
numRep2Use, 1), 
			matrix(rnorm(1*sum(numRep2Use)), sum(numRep2Use), 1))
	colnames(genData0) <- c("ID", "dose", "response", "x1")
	generatedData <- rbind(generatedData, genData0)
}
simRes <- simulEvalDRM (pilotData = 
generatedData[generatedData$ID == 2, c(2,3)], 
doseLevels = c(0, 4, 20), 
		numReplications = c(6, 3, 3), numSim = 10, 
standardDeviation = 1, EDp = 0.5,
		funcList = c("linlog", "emax", "sigEmax", "logistic"))

# }