loglik_mult: Extract log-likelihood from a mixture of multinomials

In LenaCarel/nmfem: NMF-EM Algorithm

Description

Extract log-likelihood from a mixture of multinomials

Usage

loglik_mult(X, Theta, Lambda = NULL, p)

Arguments

X	a matrix of dimension N (number of observation) x M (number of variables) containing multinomials observations.
Theta	matrix of dimension M x H.
Lambda	matrix of dimension H x K. Can be NULL.
p	vector containing the proportions of each cluster. Must be of dimension K (or H if Lambda is NULL).

Value

The function returns the log-likelihood of the data to the model

Examples

travelers <- travelers[ ,-1]
M <- ncol(travelers)
K <- 5

Theta0    <- t(dplyr::sample_n(travelers, K))
Theta0    <- Theta0 / matrix(rep(apply(Theta0, 2, sum), M), nrow = M, ncol = K, byrow = TRUE)
travelers <- as.matrix(travelers)
p0        <- rep(1 / K, K)

llh <- loglik_mult(travelers, Theta0, p = p0)
llh

LenaCarel/nmfem documentation built on Dec. 4, 2019, 12:14 p.m.