pldam: Predictive discriminant analyses based on balanced samples.

In vbonhomme/mevolCVP: Bla blah

Predictive discriminant analyses based on balanced samples.

Description

More about it

Usage

pldam(
  mat = mat,
  group = group,
  mattoid = mattoid,
  codetoid = codetoid,
  ncpts = 20,
  limproba = 0,
  limCVP = FALSE,
  nrep = 100,
  return.matrix = TRUE,
  minInd = NULL
)

Arguments

mat	matrix with the 'known' data (usually PC scores, but works also with univariate data)
group	grouping factor of the 'known' data
mattoid	matrix with the data to be predicted
codetoid	label of the specimens to predict
ncpts	number of PCs to retain
limproba	a limit for the posterior probability to be concidered
limCVP	a limit for the cross validation percent of the discriminate analysis to be concidered
nrep	number of replicates
return.matrix	logical whether to return full results
minInd	the number of specimens to include for the balanced design if smaller than the smallest of the "known" groups

Value

a list

Examples

# Some examples
# Grab some data from pig
mat <- pig$mat
gp <- pig$gp
# For the sake of reproducibility
set.seed(123)
train_ids <- sample(nrow(mat), 150, replace=FALSE)

pldam(# train data
      mat=mat[train_ids, ], group=gp[train_ids],
      # test data
      mattoid=mat[-train_ids, ], codetoid=paste0("id_", 1:nrow(mat[-train_ids, ])),
      nrep=2)

vbonhomme/mevolCVP documentation built on Aug. 24, 2024, 12:49 p.m.