DirichletMultinomial (version 1.14.0)

cvdmngroup: Cross-validation on Dirichlet-Multinomial classifiers.

Description

Run cross-validation on Dirichlet-Multinomial generative classifiers.

Usage

cvdmngroup(ncv, count, k, z, ..., verbose = FALSE, .lapply = parallel::mclapply)

Arguments

ncv

integer(1) number of cross-validation groups, between 2 and nrow(count).

count

matrix of sample x taxon counts, subsets of which are used for training and cross-validation.

k

named integer() vector of groups and number of Dirichlet components; e.g., c(Lean=1, Obese=3) performs cross-validation for models with k=1 Dirichlet components for the ‘Lean’ group, k=3 Dirichlet components for ‘Obese’.

z

True group assignment.

...

Additional arguments, passed to dmn during each cross-validation.

verbose

logical(1) indicating whether progress should be reported

.lapply

A function used to perform the outer cross-vaildation loop, e.g., lapply for calculation on a single processor, parallel::mclapply for parallel evaluation.

Value

A data.frame summarizing classifications of test samples in cross-validation groups. Columns are:

group

The cross-validation group in which the indivdual was used for testing.

additional columns

Named after classification groups, giving the posterior probability of assignment.

See Also

dmn, DirichletMultinomial-package, vignette("DirichletMultinomial")

Examples

data(xval)  ## result of following commands
head(xval)

## Not run: 
# ## count matrix
# fl <- system.file(package="DirichletMultinomial", "extdata",
#                   "Twins.csv")
# count <- t(as.matrix(read.csv(fl, row.names=1)))
# 
# ## phenotype
# fl <- system.file(package="DirichletMultinomial", "extdata",
#                   "TwinStudy.t")
# pheno0 <- scan(fl)
# lvls <- c("Lean", "Obese", "Overwt")
# pheno <- factor(lvls[pheno0 + 1], levels=lvls)
# names(pheno) <- rownames(count)
# 
# ## subset
# keep <- c("Lean", "Obese")
# count <- count[pheno 
# pheno <- factor(pheno[pheno 
# 
# ## cross-validation, single Dirichlet component for Lean, 3 for Obese
# xval <- cvdmngroup(nrow(count), count, c(Lean=1, Obese=3), pheno,
#                    verbose=TRUE, mc.preschedule=FALSE)
# ## End(Not run)