One M-Step: Compute One M-Step Based on Current Posterior Probabilities
Description
This function will perform one M-step based on current posterior
probabilities.
This is a core function of em.onestep.
Usage
m.step(PARAM)
Arguments
PARAM
a set of parameters.
Value
Returning a new PARAM which maximizes the complete data log
likelihood for the current iteration.
Details
This function will base on the current posterior probabilities
Z.spmd to estimate the parameters PARAM
mainly including mixing proportions ETA,
centers of clusters MU, and
dispersions of clusters SIGMA.