init.bmm.hyperparameters: init.bmm.hyperparameters: Provides default initialization of...

In genome/bmm: Variational Bayesian approach to fitting mixtures of beta, binomial, or Gaussian distributions

init.bmm.hyperparameters: Provides default initialization of hyperparameters for Beta mixture model.

Description

Initialize the bmm hyperparameters (to be passed to bmm or bmm.fixed.num.components)

NB: This provides what should be a generally reasonable initialization of hyperparameters. However, better results may be obtained by tuning these in an application-specific manner.

Usage

  init.bmm.hyperparameters(X, N.c)

Arguments

X	an N x D matrix with rows being the items to cluster. All entries are assumed to be proportions (i.e., between 0 and 1). Notice that there are no summation restrictions–i.e., proportions do not sum to unity across an item's dimensions.
N.c	N.c: the number of components/clusters to attempt

Value

Returns a list with the following components:

mu0	a D x N.c matrix holding the hyperparameter values of the shape parameters for the gamma prior distributions over the u parameters. i.e., mu[d,n] is the shape parameter governing u[d,n]. Introduced in eqn (15).
alpha0	a D x N.c matrix holding the hyperparameter values of the rate (i.e., inverse scale) parameters for the gamma prior distributions over the u parameters. i.e., mu[d,n] is the rate parameter governing u[d,n]. Introduced in eqn (15).
nu0	a D x N.c matrix holding the hyperparameter values of the shape parameters for the gamma prior distributions over the v parameters. i.e., nu[d,n] is the shape parameter governing v[d,n]. Introduced in eqn (16).
beta0	a D x N.c matrix holding the hyperparameter values of the rate (i.e., inverse scale) parameters for the gamma prior distributions over the v parameters. i.e., beta[d,n] is the rate parameter governing v[d,n]. Introduced in eqn (16).
c0	a vector with D components holding the hyperparameter values of the parameters of the Dirichlet distribution over the mixing coefficients pi. Introduced in eqn (19).

genome/bmm documentation built on Aug. 4, 2022, 8:01 a.m.