loglik: Loglikelihood for model-based hierarchical clustering.
Description
Gives the loglikelihood for each stage of model-based hierarchical clustering.
Usage
loglik(tree, data, ...)
Arguments
tree
an "mhtree" object.
data
the data used to produce the "mhtree" object.
...
Other arguments, depending on the method that is
appropriate. In particular Vinv, the approximate reciprocal
hypervolume of the region from which the data is drawn. The default (for
those methods that need this quantity) is determined by the
Value
the loglikelihood corresponding to the initial partition and to each stage of
merging in hierarchical clustering, together with the following attribute:
nmergethe number of clusters merged at each stage.
NOTES
The value given is equal to the loglikelihood up to an additive constant.
For those models in which they arise, indeterminate terms are
assigned the value -k*log(volume), where k is the number of observations
associated with the term.
If you scaled your data before using mhtree, be
sure to use the same scaling when supplying the data to loglik.