simula.rlg {tclust}

Simulate contaminated data set for applying rlg

Description

Simulate alpha*100% contaminated data set for applying rlg by generating a k=3 components with equal size and # common underlying dimension q_1=q_2=q_3=q

Usage

simula.rlg(q = 2, p = 10, n = 200, var = 0.01, sep.means = 0, alpha = 0.05)

Arguments

q

intrinsic dimension

p

dimension (p >= 2 and p > q)

n

number of observations

var

The smaller 'var' the smaller the scatter around the lower dimensional space

sep.means

Parameter controlling the location vectors separation

alpha

contamination level

Value

a list with the following items

• x - The generated dataset

• true - The true classification

Examples

 res <- simula.rlg(q=5, p=200, n=150, var=0.01, sep.means=0.00)
 plot(res$x,col=res$true+1)