try_mixture_gpcm {outlierMBC} | R Documentation |
Run mixture::gpcm
and try alternative covariance structures or
initialisations if necessary.
Description
If mixture::gpcm
returns an error, this function first tries
the other covariance structures, and then tries a k-means initialisation.
Usage
try_mixture_gpcm(x, comp_num, mnames, z, nmax, atol, fixed_labels)
Arguments
x |
Data. |
comp_num |
Number of mixture components. |
mnames |
Model names for mixture::gpcm. |
z |
Component assignment probability matrix for initialisation. |
nmax |
Maximum number of iterations for |
atol |
EM convergence tolerance threshold for |
fixed_labels |
Cluster labels that are known a prior. See |
Value
Object of class "gpcm"
outputted by mixture::gpcm
.
[Package outlierMBC version 0.0.1 Index]