marl (version 1.0)

negbin.rel.clust: Relative Likelihood based clustering assuming Negative Binomial distribution.

Description

The function provides weighted relative likelihood function matrix, cluster assigment based on k means clustering, and plot of weighted relative likelihood functions with colors representing cluster assignment. Each observation in the dataset is assumed to follow negative binomial distribution.

Usage

negbin.rel.clust(x, mu.min, mu.max, nclust = 2, len = 200, plot = TRUE, seed = 132)

Arguments

x

Observations of length greater than 1. Data can be entered as matrix or list.

mu.min

Minimum value of mu for the relative likelihood function.

mu.max

Maximum value of mu for the relative likelihood function.

nclust

Number of Clusters.

len

Length of values to be evaluated at in between mu.min and mu.max.

plot

If set TRUE, provides plot of weighted relative likelihood functions colored by their cluster assignment.

seed

Seed to be set for reproducibile results.

Value

Wt.Rel.Likld

A matrix of weighted relative likelihood functions with rows indicating observations. The column names are the values at which the weighted relative likelihood functions are evaluated.

Cluster.Assignment

Cluster assignment for each observation.

table

Table summarizing size of clusters.

Details

For mathematical details, please contact the authors.

References

None.

Examples

x <- sim.negbin(c(4,5,10),3,10,12)
negbin.rel.clust(x,1,20,nclust = 3,len=200,plot=TRUE,seed=132)