SpectralClMixed (version 1.0.1)

mspec: Preforms spectral clustering on mix typed data

Description

Preforms spectral clustering of mix-type data

Usage

mspec(
  z,
  k = 2,
  sigma = c(20, 20),
  c_wt = NULL,
  starts = 10,
  its = 300,
  verbose = FALSE
)

Value

A class SpectralClMixed list with components

ct_wt

the selected category weight

bt/wt_ss

the between divided by the within sum of squares

tot_wt_ss

the total within sum of squares

cluster

the cluster assignments

data

the original data

Arguments

z

data to be clustered

k

the number of clusters.

sigma

vector of lower,upper bounds for sigma

c_wt

the category weights, is assigned to c(0.9999, 0.999, .99, seq(0.95, 0.05,-0.05), .01, 0.001, 0.0001) if null.

starts

the number of random starts

its

the max number of iterations for the kmeans algorithm

verbose

if you would like printed output during running of function

References

F. Mbuga and, C. Tortora. Spectral Clustering of Mixed-Type Data. Stats, 5(1) 2022

Examples

c1=data.frame(v1=rnorm(30,0),v2=rnorm(30,0),v3=factor(round(runif(30))+1))
c2=data.frame(v1=rnorm(30,2),v2=rnorm(30,4),v3=factor(round(runif(30))+4))
data=rbind(c1,c2)
res=mspec(data, k = 2)
summary(res)
plot(res)