predict.MixtComp: Predict using RMixtComp

In RMixtComp: Mixture Models with Heterogeneous and (Partially) Missing Data

Predict using RMixtComp

Description

Predict the cluster of new samples.

Usage

## S3 method for class 'MixtComp'
predict(
  object,
  newdata = NULL,
  type = c("partition", "probabilities"),
  nClass = NULL,
  ...
)

Arguments

object	output of mixtCompLearn function.
newdata	a data.frame, a matrix or a named list containing the data (see Details and Data format sections in mixtCompLearn documentation). If NULL, use the data in object.
type	if "partition", returns the estimated partition. If "probabilities", returns the probabilities to belong to each class (tik).
nClass	the number of classes of the mixture model to use from object. If NULL, uses the number maximizing the criterion.
...	other parameters of mixtCompPredict function.

Details

This function is based on the generic method "predict". For a more complete output, use mixtCompPredict function.

Value

if type = "partition", it returns the estimated partition as a vector. If type = "probabilities", it returns the probabilities to belong to each class (tik) as a matrix.

Author(s)

Quentin Grimonprez

See Also

mixtCompPredict

Examples

data(iris)

model <- list(
  Sepal.Length = "Gaussian", Sepal.Width = "Gaussian",
  Petal.Length = "Gaussian", Petal.Width = "Gaussian"
)

resLearn <- mixtCompLearn(iris[-c(1, 51, 101), ], model = model, nClass = 1:3, nRun = 1)

# return the partition
predict(resLearn)

# return the tik for the 3 new irises for 2 and 3 classes
predict(resLearn, newdata = iris[c(1, 51, 101), ], type = "probabilities", nClass = 2)
predict(resLearn, newdata = iris[c(1, 51, 101), ], type = "probabilities", nClass = 3)

RMixtComp documentation built on July 9, 2023, 6:06 p.m.