BIC_F: computes the BIC of a RES distribution based on a finite...

In Mufabo/ICASSP20.T6.R: ICASSP20.T6.R

Description

computes the BIC of a RES distribution based on a finite sample penalty term

Usage

BIC_F(data, S_est, mu_est, t, mem, rho, psi, eta)

Arguments

data	Matrix[N, r] data samples
S_est	array[r, r, ll] Scatter matrices of the clusters
mu_est	Matrix[r, ll] estimated mean vectors of all clusters
t	Matrix[N, ll] Squared Mahalanobis distances to cluster centers
mem	Matrix[N, ll] cluster memberships
rho	Vector rho of density generator
psi	Vector. psi of density generator
eta	Vector. eta of density generator

Value

list

1. bic

2. pen : Penalty term

3. like : Likelihood term

Note

"Robust M-Estimation based Bayesian Cluster Enumeration for Real Elliptically Symmetric Distributions" Christian A. Schroth and Michael Muma, Signal Processing Group, Technische Universität Darmstadt submitted to IEEE Transactions on Signal Processing

Mufabo/ICASSP20.T6.R documentation built on May 30, 2021, 11:20 a.m.