step_fmx {QuantileGH} | R Documentation |
Forward Selection of
-parsimonious Model with Fixed Number of Components
Description
To select the -parsimonious mixture model,
i.e., with some
and/or
parameters equal to zero,
conditionally on a fixed number of components
.
Usage
step_fmx(
object,
test = c("BIC", "AIC"),
direction = c("forward", "backward"),
...
)
Arguments
object |
fmx object |
test |
character scalar, criterion to be used, either Akaike's information criterion AIC-like, or Bayesian information criterion BIC-like (default). |
direction |
character scalar, |
... |
additional parameters, currently not in use |
Details
The algorithm starts with quantile least Mahalanobis distance estimates
of either the full mixture of Tukey -&-
distributions model, or
a constrained model (i.e., some
and/or
parameters equal to zero according to the user input).
Next, each of the non-zero
and/or
parameters is tested using the likelihood ratio test.
If all tested
and/or
parameters are significantly different from zero at the level 0.05
the algorithm is stopped and the initial model is considered
-parsimonious.
Otherwise, the
or
parameter with the largest p-value is constrained to zero
for the next iteration of the algorithm.
The algorithm iterates until only significantly-different-from-zero and
parameters
are retained, which corresponds to
-parsimonious Tukey
-&-
mixture model.
Value
Function step_fmx returns an object of S3 class 'step_fmx'
,
which is a list of selected models (in reversed order) with attribute(s)
'direction'
and
'test'
.