Computes the variational expectation step of the algorithm

Computes the variational expectation step of the algorithm

VEstep(
  MO,
  SO,
  SH,
  Omega,
  W,
  Wg,
  MH,
  Pg,
  logSTW,
  logSTWg,
  alpha,
  it1,
  verbatim,
  trackJ = FALSE,
  hist = FALSE
)

Arguments

MO	Matrix of observed means.
SO	Matrix of observed marginal variances.
SH	Matrix of observed hidden variances.
Omega	matrix containing the precision terms of precision matrices faithful ot a tree.
W	Edges weights matrix.
Wg	Variational edges weights matrix.
MH	Matrix of hidden means.
Pg	Edges probabilities matrix.
logSTW	Log of the Matrix Tree quantity of the W matrix.
logSTWg	Log of the Matrix Tree quantity of the Wg matrix.
alpha	Tempering parameter.
it1	Checks if nestorFit is at its first iteration.
verbatim	Displays verbose if set to 2.
trackJ	Boolean for evaluating the lower bound at each parameter update.
hist	Boolean for printing edges weights histogram at each iteration.

Value

Quantities required by the Mstep funciton:

• Pg: edges probabilities.

• Wg: edges variational weights.

• M: variational means.

• S: variational marginal variances.

• LB: lower bound values.

• logSTWg: log of the Matrix Tree quantity of the Wg matrix.

• max.prec: boolean tracking the reach of maximal precision during computation.