{smcl}
{* 02apr2004}{...}
{hline}
help for {hi:cluster stop}{right:manual:  {hi:[CL] cluster stop}}
{right:dialog:  {dialog clstop:cluster stop}     }
{hline}

{title:Cluster analysis stopping rules}

{p 8 21 2}
	{cmd:cluster} {cmd:stop} [{it:clname}]
	[{cmd:,}
	{cmdab:r:ule:(}{cmdab:cal:inski} | {cmd:duda} | {it:rule_name}{cmd:)}
	{cmdab:gr:oups:(}{it:numlist}{cmd:)}
	{cmdab:mat:rix:(}{it:matname}{cmd:)}
	]


{title:Description}

{p 4 4 2}
The {cmd:cluster stop} command computes stopping rule values after a cluster
analysis; see help {help cluster}.  Two stopping rules are currently available
-- Calinski & Harabasz pseudo-F index and the Duda & Hart Je(2)/Je(1) index.

{p 4 4 2}{it:clname} specifies the name of the cluster analysis.  The default
is the latest performed cluster analysis, which can be reset using the
{cmd:cluster use} command; see help {help cluster utility}.


{title:Options}

{p 4 8 2}
{cmd:rule(calinski} | {cmd:duda} | {it:rule_name}{cmd:)} indicates the
stopping rule.  {cmd:rule(calinski)}, the default, specifies the Calinski &
Harabasz index.  {cmd:rule(duda)} specifies the Duda & Hart index.

{p 8 8 2}
Additional stopping rules can be added to the {cmd:cluster stop} command; see
help {help cluster programming}
and {hi:[CL] cluster programming subroutines} for details.
These additional stopping rules are then obtained using the
{cmd:rule(}{it:rule_name}{cmd:)} option.

{p 4 8 2}
{cmd:groups(}{it:numlist}{cmd:)} specifies the cluster groupings for which the
stopping rule is to be computed.  {cmd:groups(3/20)} indicates that the
measure is to be computed for the 3 group solution, the 4 group solution, ...
and the 20 group solution.

{p 8 8 2}
The default is {cmd:groups(1/15)} with {cmd:rule(duda)}.  The default is
{cmd:groups(2/15)} with {cmd:rule(calinski)} for a hierarchical cluster
analysis.  {cmd:groups(1)} is not allowed with {cmd:rule(calinski)} since the
measure is not defined for the degenerate 1 group cluster solution.  The
{cmd:groups()} option is unnecessary (and not allowed) for a nonhierarchical
cluster analysis.

{p 8 8 2}
If there are ties in the hierarchical cluster analysis structure, some (or
possibly all) of the requested stopping rule solutions may not be computable.
{bind:{cmd:cluster stop}} passes over without comment the {cmd:groups()} for
which ties in the hierarchy cause the stopping rule to be undefined.

{p 4 8 2}
{cmd:matrix(}{it:matname}{cmd:)} saves the results in a matrix named
{it:matname}.

{p 8 8 2}
With {cmd:rule(calinski)} the matrix has two columns.  The first column gives
the number of clusters and the second column gives the corresponding Calinski
& Harabasz pseudo-F stopping rule index.

{p 8 8 2}
With {cmd:rule(duda)} the matrix has three columns.  The first column gives
the number of clusters.  The second column gives the corresponding Duda & Hart
Je(2)/Je(1) stopping rule index.  The third column provides the corresponding
pseudo T-squared values.


{title:Examples}

{p 4 8 2}{cmd:. cluster stop}{p_end}
{p 4 8 2}{cmd:. cluster stop myclus, rule(duda)}{p_end}
{p 4 8 2}{cmd:. cluster stop, rule(calinski) groups(2/20) matrix(z)}


{title:Also see}

{p 4 13 2}
Manual:  {hi:[CL] cluster},{break}
{hi:[CL] cluster stop}

{p 4 13 2}
Online:  help for {help cluster};
{help cluster averagelinkage},
{help cluster centroidlinkage},
{help cluster completelinkage},
{help cluster dendrogram},
{help cluster generate},
{help cluster medianlinkage},
{help cluster notes},
{help cluster programming},
{help cluster singlelinkage},
{help cluster utility},
{help cluster wardslinkage},
{help cluster waveragelinkage},
{help dissimilarity},
{p_end}