my.normdata.gen: Multipal normal data set with three clusters generation

In SODC: Optimal Discriminant Clustering(ODC) and Sparse Optimal Discriminant Clustering(SODC)

Description

To generate multipal normal data set with three clusters

Usage

my.normdata.gen(u, p, nobs, nvar)

Arguments

u	The parameters to contral the overlap degree among clusters. The cluster centers are (-u,u),(u,u),(u,-u) respectively. The larger u, the clusters more separated
p	Total number of variables
nobs	Number of observations
nvar	Number of informative variables, nvar takes even value;

Details

To generate a multiple normal dataset with three clusters, the first nvar varialbes are informative, and the remaining p-nvar are non-informative variables. The three clusters are separated when u is large, and overlapped when u is small.

Value

x	Multiple normal dataset generated
y	The true cluster assignment for the generated x dataset

SODC documentation built on May 30, 2017, 7:11 a.m.