Description Usage Arguments Details Value Author(s) References Examples
Multiple hypothesis testing H_0 :Cβ = 0, based on composite likelihood inference for clustered data, with common covariance structure across the clusters. The critical value is based on multivariate normal quantile, using the R-package multcomp
by Hothorn, et.al.
1 2 | ComLikMulCom(data, yno , idno , y.type =c("normal", "probit", "quadratic exponential"),
type = c("many_one", "all_pair"), f)
|
data |
data set in long format containing response, covariates and one variable that identifies the cluster |
yno |
the response column number |
idno |
column number of the variable that identify the clusters |
y.type |
the model of the response variable, it could be normal, probit, or quadratic exponential |
type |
type of the test, it could be many-to-one or all-pair comparisons |
f |
the baseline variable for many to one type |
The current functions handle the clustered normal and probit data with equal size. For quadratic exponential data, clusters can have different sizes.
m |
number of individuals in each cluster, if equal |
n |
number of clusters |
p |
number of covariates |
Cov_beta |
covariance matrix of the covariates |
T_stat |
the test statistic |
normal_quantile |
the simultaneous normal quantile |
MNQ |
results of individual null hypotheses based on composite likelihood |
Bon |
results of individual null hypotheses based on Bonferroni approach |
Sidak |
results of individual null hypotheses based on Dunn-Sidak approach |
Scheffe |
results of individual null hypotheses based on Scheffe method |
Mahdis Azadbakhsh, Xin Gao, Hanna Jankowski.
Azadbakhsh,M., Gao,X., Jankowski,H. (2016) Multiple Comparisons Using Composite Likelihood in Clustered Data. Submitted.
Hothorn, T., Bretz, F., Westfall, P., Heiberger, R. M., Schutzenmeister, A.(2010) multcomp: Simultaneous Inference for General Linear Hypotheses, R package version 1.1-7, http://CRAN.R-project.org/package=multcomp
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 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 | ### artificial data set (normal)
m <- 3
n <- 20
p <- 5
v <- sample(seq(-4,4,0.05),360,replace= TRUE)
cx <- matrix(v, nrow=(n*m))
ind <- rep(0,n*m)
for(i in 1: n) {ind[(m*(i-1)+1):(m*i)]<- i}
cxx <- cbind(cx,ind)
colnames(cxx) <- c("res", "c1","c2","c3","c4", "c5", "ind")
idno <- which( colnames(cxx)=="ind" )
yno <- which( colnames(cxx)=="res" )
ComLikMulCom(cxx, yno , idno , y.type ="normal" , type = "many_one", f=3)
ComLikMulCom(cxx, yno , idno , y.type ="normal" , type = "all_pair")
### probit
v <- sample(seq(-4,4,0.05),300,replace= TRUE)
y <- sample(c(0,1),60,replace= TRUE)
cx <- matrix(v, nrow=60)
colnames(cx) <- c("c1","c2","c3","c4", "c5")
id <- rep(0,n*m)
for(i in 1: n) {id[(m*(i-1)+1):(m*i)]<- i}
cxx <- as.data.frame(cbind(y,cx,id))
idno <- which( colnames(cxx)=="id" )
yno <- which( colnames(cxx)=="y" )
ComLikMulCom(cxx, yno , idno , y.type ="probit" , type = "all_pair")
### Normal
m <- 3
n <- 40
p <- 5
library(MASS)
library(mvtnorm)
beta <- c(0,2,0,0,1)
sigma <- matrix(c(0.8, 0.4, 0.4, 0.4, 0.8, 0.4, 0.4, 0.4, 0.8), nrow=3)
desx <- array(dim=c(m,p,n))
y <- matrix(0,nrow=m,ncol=n)
mu <- matrix(0,nrow=m,ncol=n)
for (k in 1:n){
for(j in 1:p){
desx[,j,k] <- rnorm(m,1,3.5)
}
mu[,k] <- desx[,,k]
y[,k] <- mvrnorm(1,mu[,k], sigma, tol = 1e-6, empirical = FALSE)
}
sss <- array(dim=c(m,(p+1),n))
for(i in 1:n){
ind <- rep(i,m)
sss[,,i] <- cbind(ind,desx[,,i])
}
yyy <- rep(0,120)
for( i in 1:40){
yyy[((i-1)*3+1):(i*3)] <- y[,i]
}
xxx <- matrix(0,nrow=120, ncol=6)
for( i in 1:40){
xxx[((i-1)*3+1):(i*3),] <- sss[,,i]
}
dat <- cbind(yyy,xxx)
colnames(dat) <- c("response","id", "c1","c2","c3","c4", "c5")
idno <- which( colnames(dat)=="id" )
yno <- which( colnames(dat)=="response" )
ComLikMulCom(dat, yno , idno , y.type ="normal" , type = "many_one", f=3)
ComLikMulCom(dat, yno , idno , y.type ="normal" , type = "all_pair")
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.