Skip to contents

Cluster spherical observations using a mixture of Poisson kernel-based densities

Source: R/clustering_functions.R
predict.pkbc.Rd

Obtain predictions of membership for spherical observations based on a mixture of Poisson kernel-based densities estimated by pkbc

Usage

# S4 method for class 'pkbc'
predict(object, k, newdata = NULL)

Arguments

object

Object of class pkbc

k

Number of clusters to be used.

newdata

a data.frame or a matrix of the data. If missing the clustering data obtained from the pkbc object are classified.

Value

Returns a list with the following components

  • Memb: vector of predicted memberships of newdata

  • Probs: matrix where entry (i,j) denotes the probability that observation i belongs to the k-th cluster.

See also

pkbc() for the clustering algorithm
pkbc for the class object definition.

Examples

# generate data
dat <- rbind(matrix(rnorm(100), ncol = 2), matrix(rnorm(100, 5), ncol = 2))
res <- pkbc(dat, 2)

# extract membership of dat
predict(res, k = 2)
#>   [1] 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1
#>  [38] 1 1 2 2 1 1 1 1 1 2 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
#>  [75] 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
# predict membership of new data
newdat <- rbind(matrix(rnorm(10), ncol = 2), matrix(rnorm(10, 5), ncol = 2))
predict(res, k = 2, newdat)
#> $Memb
#>  [1] 2 1 1 1 1 2 2 2 2 2
#> 
#> $Probs
#>             [,1]       [,2]
#>  [1,] 0.19565742 0.80434258
#>  [2,] 0.96607103 0.03392897
#>  [3,] 0.94966743 0.05033257
#>  [4,] 0.63318436 0.36681564
#>  [5,] 0.95972637 0.04027363
#>  [6,] 0.09664861 0.90335139
#>  [7,] 0.03520711 0.96479289
#>  [8,] 0.16335367 0.83664633
#>  [9,] 0.03340081 0.96659919
#> [10,] 0.06562115 0.93437885
#>