Distance to centroid classifier function

Given an \(n x m\) matrix of centroids, where \(m\) are the prototypic centroids with \(n\) features, classify new samples according to the distance to the centroids.

cluster_classify(data, centroid, method = "pearson")

Arguments

data	a data.frame of dimensions \(n x p\) with the samples to classify, were \(n\) are the same set of features as in the centroids
centroid	a data.frame of dimensions \(n x m\), where each column is a prototypic centroid to classify the samples
method	Character string indicating which method to use to calculate distance to centroid. Options are "pearson" (default), "kendall", or "spearman"

Value

Returns a numeric vector of length \(p\) with the class assigned to each sample according to the shortest distance to centroid

Examples

# load example dataset
require(iC10TrainingData)
require(pamr)

data(train.Exp)
data(IntClustMemb)
TrainData <- list(x = train.Exp, y = IntClustMemb)

# Create prototypic centroids
pam <- pamr.train(TrainData)
#> 123456789101112131415161718192021222324252627282930
centroids <- pam$centroids

Class <- cluster_classify(train.Exp, centroids)
table(Class, IntClustMemb)
#>      IntClustMemb
#> Class   1   2   3   4   5   6   7   8   9  10
#>    1   69   0   5   7   1   0   4   0   0   1
#>    2    0  40   2   2   0   0   0   0   1   0
#>    3    1   1 114   5   1   1   0  10   0   1
#>    4    0   1   2 110   1   0   1   0   0   0
#>    5    1   1   0   8  84   0   1   0   1   0
#>    6    0   0   1   3   0  42   3   0   0   0
#>    7    1   1   0  12   0   0  95   2   3   0
#>    8    0   0  27   0   1   0   1 129   1   0
#>    9    0   1   2   5   2   1   3   1  56   0
#>    10   4   0   3  15   4   0   1   1   5  94