predict: Extend Clustering to New Observations

In suren-rathnayake/EMMIXmfa: Mixture Models with Component-Wise Factor Analyzers

Description

Given a fitted model of class 'emmix' (or of class 'mfa', 'mcfa', 'mtfa' and 'mctfa'), the predict function predict clusters for observations.

Usage

## S3 method for class 'emmix'
predict(object, Y, ...)

Arguments

object	An object of class 'emmix'.
Y	A data matrix with variable in the same column locations as the data used in fitting the model object.
...	Not used.

Details

A vector integers of length equal to number of observations (rows) in the data. The integers range from 1 to g where g in the number of components in the model.

The variables in Y of the predict function should be in the order as those used in obtaining the fitted model object.

Examples

set.seed(42)
test <- sample(1 : nrow(iris), 100)
model <- mfa(iris[test, -5], g=3, q=3, itmax=500, nkmeans=3, nrandom=5)
pred_clust <- predict(model, iris[-test, -5])
minmis(pred_clust, iris[-test, 5])

suren-rathnayake/EMMIXmfa documentation built on May 17, 2019, 8:45 p.m.