Description Usage Arguments Value Examples
This function performs a multiple contrast test. Optimal contrasts are derived based on the trend you choose.
1 2 3 4 |
muHat |
a numeric vector of expected values in all doses groups including placebo. |
S |
convariance matrix. |
df |
specify degree freedom for multivariate t. If this argument is missing, df=0 corresponding multivariate normal. |
doses |
a numeric vector of specified doses. When this argument is missing, doses are the object in the specified model if model is given for MCP Mod trend test, or doses are 0 to k by 1 for all other trend test. |
alternative |
a single character string, specifying the alternative for the multiple contrast trend test. If this argument is missing, alternative="one.sided". |
trend |
a single character string, determining the trend for the multiple contrast trend test, one of "CPL", "MCP Mod","Aug Williams", "Williams", "Dunnett", "Linear","Arbitrary". Note that when "MCP_Mod" is specified for trend, model needs to be specified. When "Arbitrary" is specified, muMat needs to be specified. |
model |
specify models if you choose trend to be MCP Mod trend test. |
muMat |
a numeric vector or matrix specified for mu when trend is specified to be "Arbitrary". |
doses, optimal contrast, correlation contrast, test statisticis, p values
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 | load(DoseFinding)
transformDoses<-function(doses,mindose,mult) {
log(1+(mult/mindose)*doses)
}
itransDoses<-function(doses,mindose,mult) {
(exp(doses)-1)*mindose/mult
}
origDose<-c(0,2.5,5,10,20,50,100,200)
transDose<-transformDoses(doses=origDose,mindose=2.5,mult=5)
dat<-data.frame(dose=as.factor(origDose),n=c(133,32,44,63,63,65,59,58),
respn=c(13,4,5,16,12,14,14,21))
dat$resprate<-dat$respn/dat$n
dat$txeff<-dat$resprate-dat$resprate[1]
logfit<-glm(cbind(respn,n-respn)~dose+0,family=binomial,data=dat)
muHat<-coef(logfit)
S<-vcov(logfit)
model<-Mods(
logistic=c(log(1+2*10),0.5),
sigEmax=c(log(1+2*10),4),
emax=c(log(1+2*0.1),log(1+2*1.0)),
linear=NULL,
exponential=log(1+2*1.0),
quadratic= -1/log(1+2*550),
doses=transDose)
test_MCP<-trendtesting(muHat=muHat, S=S, doses=transDose, alternative="one.sided", trend="MCP Mod", model=model)
#with arbitrary trend
#arbitrary mu
mu<-mu.linear(7, doses=seq(0,7,1))
test_MCP<-trendtesting(muHat=muHat, S=S, doses=transDose, alternative="one.sided", trend="Arbitrary", muMat=mu)
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.