| Title: | Multiple Imputation for Multivariate Data |
|---|---|
| Description: | Fully parametric Bayesian multiple imputation framework for massive multivariate data of different variable types as seen in Demirtas, H. (2017) <doi:10.1007/978-981-10-3307-0_8>. |
| Authors: | Rawan Allozi, Hakan Demirtas |
| Maintainer: | Rawan Allozi <[email protected]> |
| License: | GPL-2 | GPL-3 |
| Version: | 1.0 |
| Built: | 2024-12-01 07:58:56 UTC |
| Source: | CRAN |
This package implements a Bayesian multiple imputation framework for multivariate data. Most incomplete data sets constist of interdependent binary, ordinal, count, and continuous data. Furthermore, planned missing data designs have been developed to reduce respondent burden and lower the cost associated with data collection. The unified, general-purpose multiple imputation framework described in Demirtas (2017) can be utilized in developing power analysis guidelines for intensive multivariate data sets that are collected via increasingly popular real-time data capture (RTDC) approaches. This framework can accommodate all four major types of variables with a minimal set of assumptions. The data are prepared for multivariate normal multiple imputation for use in the norm package and subsequently backtransformed to the original distribution.
This package consists of one main function and six auxiliary functions. Multiple imputation can be performed using the function MI. While the auxiliary functions are utilized in MI, they can be used as stand-alone functions. nctsum outputs a list with summary statistics and Fleishman coefficients and standardized forms of each variable. ordmps is utilized for ordinal variables and outputs a list with empirical marginal probabilities and the associated observations for each ordinal variable. countrate is designed for variables and outputs a list with empirical rates and the associated observations for each ordinal variable. MVN.corr calculates the intermediate correlation matrix, MVN.dat transforms variables to a standard normal variable, and trMVN.dat transforms standard normal variables to ordinal, count, and/or non-normal continuous variables through specified parameters.
| Package: | MultiVarMI |
| Type: | Package |
| Version: | 1.0 |
| Date: | 2018-04-08 |
| License: | GPL-2 | GPL-3 |
Rawan Allozi, Hakan Demirtas
Maintainer: Rawan Allozi <[email protected]>
Demirtas, H. and Hedeker, D. (2011). A practical way for computing approximate lower and upper correlation bounds. The American Statistician, 65(2), 104-109.
Demirtas, H., Hedeker, D., and Mermelstein, R. J. (2012). Simulation of massive public health data by power polynomials. Statistics in Medicine, 31(27), 3337-3346.
Demirtas, H. (2016). A note on the relationship between the phi coefficient and the tetrachoric correlation under nonnormal underlying distributions. The American Statistician, 70(2), 143-148.
Demirtas, H. and Hedeker, D. (2016). Computing the point-biserial correlation under any underlying continuous distribution. Communications in Statistics-Simulation and Computation, 45(8), 2744-2751.
Demirtas, H., Ahmadian, R., Atis, S., Can, F.E., and Ercan, I. (2016). A nonnormal look at polychoric correlations: modeling the change in correlations before and after discretization. Computational Statistics, 31(4), 1385-1401.
Demirtas, H. (2017). A multiple imputation framework for massive multivariate data of different variable types: A Monte-Carlo technique. Monte-Carlo Simulation-Based Statistical Modeling, edited by Ding-Geng (Din) Chen and John Dean Chen, Springer, 143-162.
Ferrari, P.A. and Barbiero, A. (2012). Simulating ordinal data. Multivariate Behavioral Research, 47(4), 566-589.
Fleishman A.I. (1978). A method for simulating non-normal distributions. Psychometrika, 43(4), 521-532.
Vale, C.D. and Maurelli, V.A. (1983). Simulating multivariate nonnormal distributions. Psychometrika, 48(3), 465-471.
This function computes the empirical rates for count data.
countrate(count.dat)countrate(count.dat)
count.dat |
A matrix consisting of count variables. |
A list of length ncol(count.dat) containing the data and empirical rates for each variable in count.dat.
library(PoisBinOrdNonNor) set.seed(123) n<-1e4 lambdas<-list(1, 3) #generate Poisson data cmat.star <- find.cor.mat.star(cor.mat = .4 * diag(2) + .6, no.pois = length(lambdas), pois.list = lambdas) cntdata <- genPBONN(n, no.pois = length(lambdas), cmat.star = cmat.star, pois.list = lambdas) #set a sample of the data to missing cntdata<-apply(cntdata, 2, function(x) { x[sample(1:n, size=n/10)]<-NA return(x) }) cntdata<-data.frame(cntdata) cntinfo<-countrate(cntdata)library(PoisBinOrdNonNor) set.seed(123) n<-1e4 lambdas<-list(1, 3) #generate Poisson data cmat.star <- find.cor.mat.star(cor.mat = .4 * diag(2) + .6, no.pois = length(lambdas), pois.list = lambdas) cntdata <- genPBONN(n, no.pois = length(lambdas), cmat.star = cmat.star, pois.list = lambdas) #set a sample of the data to missing cntdata<-apply(cntdata, 2, function(x) { x[sample(1:n, size=n/10)]<-NA return(x) }) cntdata<-data.frame(cntdata) cntinfo<-countrate(cntdata)
This function implements the multiple imputation framework as described in Demirtas (2017) "A multiple imputation framework for massive multivariate data of different variable types: A Monte-Carlo technique."
MI(dat, var.types, m)MI(dat, var.types, m)
dat |
A data frame containing multivariate data with missing values. |
var.types |
The variable type corresponding to each column in dat, taking values of "NCT" for continuous data, "O" for ordinal or binary data, or "C" for count data. |
m |
The number of stochastic simulations in which the missing values are replaced. |
A list containing m imputed data sets.
Demirtas, H. and Hedeker, D. (2011). A practical way for computing approximate lower and upper correlation bounds. The American Statistician, 65(2), 104-109.
Demirtas, H., Hedeker, D., and Mermelstein, R. J. (2012). Simulation of massive public health data by power polynomials. Statistics in Medicine, 31(27), 3337-3346.
Demirtas, H. (2016). A note on the relationship between the phi coefficient and the tetrachoric correlation under nonnormal underlying distributions. The American Statistician, 70(2), 143-148.
Demirtas, H. and Hedeker, D. (2016). Computing the point-biserial correlation under any underlying continuous distribution. Communications in Statistics-Simulation and Computation, 45(8), 2744-2751.
Demirtas, H., Ahmadian, R., Atis, S., Can, F.E., and Ercan, I. (2016). A nonnormal look at polychoric correlations: modeling the change in correlations before and after discretization. Computational Statistics, 31(4), 1385-1401.
Demirtas, H. (2017). A multiple imputation framework for massive multivariate data of different variable types: A Monte-Carlo technique. Monte-Carlo Simulation-Based Statistical Modeling, edited by Ding-Geng (Din) Chen and John Dean Chen, Springer, 143-162.
Ferrari, P.A. and Barbiero, A. (2012). Simulating ordinal data. Multivariate Behavioral Research, 47(4), 566-589.
Fleishman A.I. (1978). A method for simulating non-normal distributions. Psychometrika, 43(4), 521-532.
Vale, C.D. and Maurelli, V.A. (1983). Simulating multivariate nonnormal distributions. Psychometrika, 48(3), 465-471.
library(PoisBinOrdNonNor) set.seed(1234) n<-1e5 lambdas<-list(1, 3) #2 count variables mps<-list(c(.2, .8), c(.6, 0, .3, .1)) #1 binary variable, 1 ordinal variable with skip pattern moms<-list(c(-1, 1, 0, 1), c(0, 3, 0, 2)) #2 continuous variables ################################################ #Generate Poisson, Ordinal, and Continuous Data# ################################################ #get intermediate correlation matrix cmat.star <- find.cor.mat.star(cor.mat = .8 * diag(6) + .2, #all pairwise correlations set to 0.2 no.pois = length(lambdas), no.ord = length(mps), no.nonn = length(moms), pois.list = lambdas, ord.list = mps, nonn.list = moms) #generate dataset mydata <- genPBONN(n, no.pois = length(lambdas), no.ord = length(mps), no.nonn = length(moms), cmat.star = cmat.star, pois.list = lambdas, ord.list = mps, nonn.list = moms) cor(mydata) apply(mydata, 2, mean) #Make 10 percent of each variable missing completely at random mydata<-apply(mydata, 2, function(x) { x[sample(1:n, size=n*0.1)]<-NA return(x) } ) #Create 5 imputed datasets mydata<-data.frame(mydata) mymidata<-MI(dat=mydata, var.types=c('C', 'C', 'O', 'O', 'NCT', 'NCT'), m=5) #get the means of each variable for the m imputed datasets do.call(rbind, lapply(mymidata, function(x) apply(x, 2, mean))) #get m correlation matrices of for the m imputed dataset lapply(mymidata, function(x) cor(x)) #Look at the second imputed dataset head(mymidata$dataset2) ##run a linear model on each dataset and extract coefficients mycoef<-lapply(mymidata, function(x) { fit<-lm(X6~., data=data.frame(x)) fit.coef<-coef(fit) return(fit.coef) }) do.call(rbind, mycoef)library(PoisBinOrdNonNor) set.seed(1234) n<-1e5 lambdas<-list(1, 3) #2 count variables mps<-list(c(.2, .8), c(.6, 0, .3, .1)) #1 binary variable, 1 ordinal variable with skip pattern moms<-list(c(-1, 1, 0, 1), c(0, 3, 0, 2)) #2 continuous variables ################################################ #Generate Poisson, Ordinal, and Continuous Data# ################################################ #get intermediate correlation matrix cmat.star <- find.cor.mat.star(cor.mat = .8 * diag(6) + .2, #all pairwise correlations set to 0.2 no.pois = length(lambdas), no.ord = length(mps), no.nonn = length(moms), pois.list = lambdas, ord.list = mps, nonn.list = moms) #generate dataset mydata <- genPBONN(n, no.pois = length(lambdas), no.ord = length(mps), no.nonn = length(moms), cmat.star = cmat.star, pois.list = lambdas, ord.list = mps, nonn.list = moms) cor(mydata) apply(mydata, 2, mean) #Make 10 percent of each variable missing completely at random mydata<-apply(mydata, 2, function(x) { x[sample(1:n, size=n*0.1)]<-NA return(x) } ) #Create 5 imputed datasets mydata<-data.frame(mydata) mymidata<-MI(dat=mydata, var.types=c('C', 'C', 'O', 'O', 'NCT', 'NCT'), m=5) #get the means of each variable for the m imputed datasets do.call(rbind, lapply(mymidata, function(x) apply(x, 2, mean))) #get m correlation matrices of for the m imputed dataset lapply(mymidata, function(x) cor(x)) #Look at the second imputed dataset head(mymidata$dataset2) ##run a linear model on each dataset and extract coefficients mycoef<-lapply(mymidata, function(x) { fit<-lm(X6~., data=data.frame(x)) fit.coef<-coef(fit) return(fit.coef) }) do.call(rbind, mycoef)
This function calculates an intermediate correlation matrix for Poisson, ordinal, and continuous random variables, with specified target correlations and marginal properties.
MVN.corr(indat, var.types, ord.mps=NULL, nct.sum=NULL, count.rate=NULL)MVN.corr(indat, var.types, ord.mps=NULL, nct.sum=NULL, count.rate=NULL)
indat |
A data frame containing multivariate data. Continuous variables should be standardized. |
var.types |
The variable type corresponding to each column in dat, taking values of "NCT" for continuous data, "O" for ordinal or binary data, or "C" for count data. |
ord.mps |
A list containing marginal probabilities for binary and ordinal variables as packaged from output in |
nct.sum |
A matrix containing summary statistics for continuous variables as packaged from output in |
count.rate |
A vector containing rates for count variables as packaged from output in |
The intermediate correlation matrix.
Demirtas, H. and Hedeker, D. (2011). A practical way for computing approximate lower and upper correlation bounds. The American Statistician, 65(2), 104-109.
Demirtas, H., Hedeker, D., and Mermelstein, R. J. (2012). Simulation of massive public health data by power polynomials. Statistics in Medicine, 31(27), 3337-3346.
Demirtas, H. (2016). A note on the relationship between the phi coefficient and the tetrachoric correlation under nonnormal underlying distributions. The American Statistician, 70(2), 143-148.
Demirtas, H. and Hedeker, D. (2016). Computing the point-biserial correlation under any underlying continuous distribution. Communications in Statistics-Simulation and Computation, 45(8), 2744-2751.
Demirtas, H., Ahmadian, R., Atis, S., Can, F.E., and Ercan, I. (2016). A nonnormal look at polychoric correlations: modeling the change in correlations before and after discretization. Computational Statistics, 31(4), 1385-1401.
Ferrari, P.A. and Barbiero, A. (2012). Simulating ordinal data. Multivariate Behavioral Research, 47(4), 566-589.
Fleishman A.I. (1978). A method for simulating non-normal distributions. Psychometrika, 43(4), 521-532.
Vale, C.D. and Maurelli, V.A. (1983). Simulating multivariate nonnormal distributions. Psychometrika, 48(3), 465-471.
MI, MVN.dat, ordmps, nctsum, countrate
library(PoisBinOrdNonNor) n<-1e4 lambdas<-list(1, 3) mps<-list(c(.2, .8), c(.6, 0, .3, .1)) moms<-list(c(-1, 1, 0, 1), c(0, 3, 0, 2)) #generate Poisson, ordinal, and continuous data cmat.star <- find.cor.mat.star(cor.mat = .8 * diag(6) + .2, no.pois = length(lambdas), no.ord = length(mps), no.nonn = length(moms), pois.list = lambdas, ord.list = mps, nonn.list = moms) mydata <- genPBONN(n, no.pois = length(lambdas), no.ord = length(mps), no.nonn = length(moms), cmat.star = cmat.star, pois.list = lambdas, ord.list = mps, nonn.list = moms) #set a sample of each variable to missing mydata<-apply(mydata, 2, function(x) { x[sample(1:n, size=n/10)]<-NA return(x) }) mydata<-data.frame(mydata) #get information for use in function ord.info<-ordmps(ord.dat=mydata[,c('X3', 'X4')]) nct.info<-nctsum(nct.dat=mydata[,c('X5', 'X6')]) count.info<-countrate(count.dat=mydata[,c('X1', 'X2')]) #extract marginal probabilites, continuous properties, and count rates mps<-sapply(ord.info, "[[", 2) nctsum<-sapply(nct.info, "[[", 2) rates<-sapply(count.info, "[[", 2) #replace continuous with standardized forms mydata[,c('X5', 'X6')]<-sapply(nct.info, "[[", 1)[,c('X5', 'X6')] var.types<-c('C', 'C', 'O', 'O', 'NCT', 'NCT') mvn.cmat<-MVN.corr(indat=mydata, var.types=var.types, ord.mps=mps, nct.sum=nctsum, count.rate=rates)library(PoisBinOrdNonNor) n<-1e4 lambdas<-list(1, 3) mps<-list(c(.2, .8), c(.6, 0, .3, .1)) moms<-list(c(-1, 1, 0, 1), c(0, 3, 0, 2)) #generate Poisson, ordinal, and continuous data cmat.star <- find.cor.mat.star(cor.mat = .8 * diag(6) + .2, no.pois = length(lambdas), no.ord = length(mps), no.nonn = length(moms), pois.list = lambdas, ord.list = mps, nonn.list = moms) mydata <- genPBONN(n, no.pois = length(lambdas), no.ord = length(mps), no.nonn = length(moms), cmat.star = cmat.star, pois.list = lambdas, ord.list = mps, nonn.list = moms) #set a sample of each variable to missing mydata<-apply(mydata, 2, function(x) { x[sample(1:n, size=n/10)]<-NA return(x) }) mydata<-data.frame(mydata) #get information for use in function ord.info<-ordmps(ord.dat=mydata[,c('X3', 'X4')]) nct.info<-nctsum(nct.dat=mydata[,c('X5', 'X6')]) count.info<-countrate(count.dat=mydata[,c('X1', 'X2')]) #extract marginal probabilites, continuous properties, and count rates mps<-sapply(ord.info, "[[", 2) nctsum<-sapply(nct.info, "[[", 2) rates<-sapply(count.info, "[[", 2) #replace continuous with standardized forms mydata[,c('X5', 'X6')]<-sapply(nct.info, "[[", 1)[,c('X5', 'X6')] var.types<-c('C', 'C', 'O', 'O', 'NCT', 'NCT') mvn.cmat<-MVN.corr(indat=mydata, var.types=var.types, ord.mps=mps, nct.sum=nctsum, count.rate=rates)
This function assigns a normal score to binary and ordinal variables using normal quantiles in this appropriate range dictated by marginal proportions; a normal score to count variables based on the equivalence of CDFs of Poisson and normal distribution in the appropriate range dictated by the rate parameters; and a normal score for each continuous measurement by finding the normal root in the Fleishman equation.
MVN.dat(ord.info=NULL, nct.info=NULL, count.info=NULL)MVN.dat(ord.info=NULL, nct.info=NULL, count.info=NULL)
ord.info |
A list containing binary and ordinal data and corresponding marginal probabilities as packaged in |
nct.info |
A list containing standardized continuous data and corresponding summary statistics for continuous variables as packaged in |
count.info |
A list containing count data and corresponding rates as packaged in |
A matrix containing normal scores for each variable input.
Fleishman A.I. (1978). A method for simulating non-normal distributions. Psychometrika, 43(4), 521-532.
library(PoisBinOrdNonNor) n<-1e4 lambdas<-list(1) mps<-list(c(.2, .8)) moms<-list(c(-1, 1, 0, 1)) #generate Poisson, ordinal, and continuous data cmat.star <- find.cor.mat.star(cor.mat = .8 * diag(3) + .2, no.pois = length(lambdas), no.ord = length(mps), no.nonn = length(moms), pois.list = lambdas, ord.list = mps, nonn.list = moms) mydata <- genPBONN(n, no.pois = length(lambdas), no.ord = length(mps), no.nonn = length(moms), cmat.star = cmat.star, pois.list = lambdas, ord.list = mps, nonn.list = moms) #set a sample of each variable to missing mydata<-apply(mydata, 2, function(x) { x[sample(1:n, size=n/10)]<-NA return(x) }) mydata<-data.frame(mydata) #get information for use in function count.info<-countrate(count.dat=data.frame(mydata[,c('X1')])) ord.info<-ordmps(ord.dat=data.frame(mydata[,c('X2')])) nct.info<-nctsum(nct.dat=data.frame(mydata[,c('X3')])) mvn.dat<-MVN.dat(ord.info=ord.info, nct.info=nct.info, count.info=count.info) #outputs in order of continuous, ordinal, countlibrary(PoisBinOrdNonNor) n<-1e4 lambdas<-list(1) mps<-list(c(.2, .8)) moms<-list(c(-1, 1, 0, 1)) #generate Poisson, ordinal, and continuous data cmat.star <- find.cor.mat.star(cor.mat = .8 * diag(3) + .2, no.pois = length(lambdas), no.ord = length(mps), no.nonn = length(moms), pois.list = lambdas, ord.list = mps, nonn.list = moms) mydata <- genPBONN(n, no.pois = length(lambdas), no.ord = length(mps), no.nonn = length(moms), cmat.star = cmat.star, pois.list = lambdas, ord.list = mps, nonn.list = moms) #set a sample of each variable to missing mydata<-apply(mydata, 2, function(x) { x[sample(1:n, size=n/10)]<-NA return(x) }) mydata<-data.frame(mydata) #get information for use in function count.info<-countrate(count.dat=data.frame(mydata[,c('X1')])) ord.info<-ordmps(ord.dat=data.frame(mydata[,c('X2')])) nct.info<-nctsum(nct.dat=data.frame(mydata[,c('X3')])) mvn.dat<-MVN.dat(ord.info=ord.info, nct.info=nct.info, count.info=count.info) #outputs in order of continuous, ordinal, count
This function calculates mean, variance, skewness, excess kurtosis, and Fleishman coefficients for continuous data and also standardizes each variable.
nctsum(nct.dat)nctsum(nct.dat)
nct.dat |
A data frame consisting of continuous variables. |
A list of length ncol(nct.dat) containing the standardized data and summary statistics for each variable in nct.dat.
Fleishman A.I. (1978). A method for simulating non-normal distributions. Psychometrika, 43(4), 521-532.
library(PoisBinOrdNonNor) set.seed(123) n<-1e4 #first four moments for each simulated variable moms<-list(c(-1, 1, 0, 1), c(0, 3, 0, 2)) #generate continuous data cmat.star <- find.cor.mat.star(cor.mat = .8 * diag(2) + .2, no.nonn = 2, nonn.list = moms) nctdata <- genPBONN(n, no.nonn = length(moms), cmat.star = cmat.star, nonn.list = moms) #set a sample of each variable to missing nctdata<-apply(nctdata, 2, function(x) { x[sample(1:n, size=n/10)]<-NA return(x) }) nctdata<-data.frame(nctdata) nctinfo<-nctsum(nctdata)library(PoisBinOrdNonNor) set.seed(123) n<-1e4 #first four moments for each simulated variable moms<-list(c(-1, 1, 0, 1), c(0, 3, 0, 2)) #generate continuous data cmat.star <- find.cor.mat.star(cor.mat = .8 * diag(2) + .2, no.nonn = 2, nonn.list = moms) nctdata <- genPBONN(n, no.nonn = length(moms), cmat.star = cmat.star, nonn.list = moms) #set a sample of each variable to missing nctdata<-apply(nctdata, 2, function(x) { x[sample(1:n, size=n/10)]<-NA return(x) }) nctdata<-data.frame(nctdata) nctinfo<-nctsum(nctdata)
This function computes the empirical marginal probabilities for binary and ordinal data.
ordmps(ord.dat)ordmps(ord.dat)
ord.dat |
A data frame consisting of binary and ordinal variables. |
A list of length ncol(ord.dat) containing the data and empirical marginal probabilities for each variable in ord.dat.
library(PoisBinOrdNonNor) set.seed(123) n<-1e4 mps<-list(c(.2, .8), c(.6, 0, .3, .1)) #generate ordinal data cmat.star <- find.cor.mat.star(cor.mat = .8 * diag(2) + .2, no.ord = length(mps), ord.list = mps) orddata <- genPBONN(n, no.ord = length(mps), cmat.star = cmat.star, ord.list = mps) #set a sample of each variable to missing orddata<-apply(orddata, 2, function(x) { x[sample(1:n, size=n/10)]<-NA return(x) }) orddata<-data.frame(orddata) ordinfo<-ordmps(orddata)library(PoisBinOrdNonNor) set.seed(123) n<-1e4 mps<-list(c(.2, .8), c(.6, 0, .3, .1)) #generate ordinal data cmat.star <- find.cor.mat.star(cor.mat = .8 * diag(2) + .2, no.ord = length(mps), ord.list = mps) orddata <- genPBONN(n, no.ord = length(mps), cmat.star = cmat.star, ord.list = mps) #set a sample of each variable to missing orddata<-apply(orddata, 2, function(x) { x[sample(1:n, size=n/10)]<-NA return(x) }) orddata<-data.frame(orddata) ordinfo<-ordmps(orddata)
This function backtransforms normal scores for ordinal variables using the thresholds determined by the marginal proportions using quantiles of the normal distribution; normal scores for continuous variables by the sum of linear combinations of standard normals using the corresponding Fleishman coefficients; and normal scores for count variables by the inverse cdf matching procedure.
trMVN.dat(indat, ord.mps=NULL, nct.sum=NULL, count.rate=NULL)trMVN.dat(indat, ord.mps=NULL, nct.sum=NULL, count.rate=NULL)
indat |
A list of data frames of normal scores to be backtransformed. |
ord.mps |
A list containing marginal probabilities for binary and ordinal variables as packaged from output in |
nct.sum |
A matrix containing summary statistics for continuous variables as packaged from output in |
count.rate |
A vector containing rates for count variables as packaged from output in |
A list containing backtransformed data.
Fleishman A.I. (1978). A method for simulating non-normal distributions. Psychometrika, 43(4), 521-532.
sndat<-data.frame(matrix(rnorm(1e4), ncol=5, nrow=1e4/5)) #ordinal marginal probabilities m1<-c(0.4, 0.6) names(m1)<-c(0,1) m2<-c(0.2, 0.3, 0.5) names(m2)<-c(0,2,3) mps<-list(X1=m1, X2=m2) #count rates rates<-c(2, 3) names(rates)<-c('X3', 'X4') #continuous nctsum<-data.frame(X5=c(1, 1, -0.31375, 0.82632, 0.31375, 0.02271)) #Weibull(1,1) rownames(nctsum)<-c('Mean', 'Variance', 'a', 'b', 'c', 'd') trdat<-trMVN.dat(indat=list(sndat), ord.mps=mps, nct.sum=nctsum, count.rate=rates)sndat<-data.frame(matrix(rnorm(1e4), ncol=5, nrow=1e4/5)) #ordinal marginal probabilities m1<-c(0.4, 0.6) names(m1)<-c(0,1) m2<-c(0.2, 0.3, 0.5) names(m2)<-c(0,2,3) mps<-list(X1=m1, X2=m2) #count rates rates<-c(2, 3) names(rates)<-c('X3', 'X4') #continuous nctsum<-data.frame(X5=c(1, 1, -0.31375, 0.82632, 0.31375, 0.02271)) #Weibull(1,1) rownames(nctsum)<-c('Mean', 'Variance', 'a', 'b', 'c', 'd') trdat<-trMVN.dat(indat=list(sndat), ord.mps=mps, nct.sum=nctsum, count.rate=rates)