robustINIT {RMBC}

robustINIT

Description

Robust Initializer for RMBC algorithm, it depends on the package ktaucenters

Usage

robustINIT(Y, K, nstart = 10)

Arguments

Y

A matrix of size n x p.

K

The number of groups

nstart

the number of starting points to the algorithm, defaults to 10

Value

A list including the initial parameters of the mixture distribution, namely

• alphaINIT: K numeric values representing the convex combination coefficients.

• muINIT: a list of length K with the location initial estimators.

• sigmaINIT: a list of length K with the scatter matrix estimators.

• indicesINIT: indices with initial clusters