mds_postestimation.hlp

是一个经济学管理应用软件 很难找的 但是经济学学生又必须用到

{smcl}
{* 03may2005}{...}
{cmd:help mds postestimation} {...}
{right:dialogs:  {bf:{dialog mds_p:predict}  {dialog mds_estat:estat}}{space 7}}
{right: {bf:{dialog mdsconfig}  {dialog mdsshepard}}}
{right:  also see:  {bf:{help mds}, {help mdslong}, {help mdsmat}} }
{hline}

{title:Title}

{p 4 33 2}
{hi:[MV] mds postestimation} {hline 2}
Postestimation tools for {cmd:mds}, {cmd:mdsmat}, and {cmd:mdslong}


{title:Description}

{pstd}
The following postestimation commands are of special interest after {cmd:mds},
{cmd:mdsmat}, and {cmd:mdslong}.

{synoptset 22 tabbed}{...}
{p2coldent:command}description{p_end}
{synoptline}
{synopt:{helpb mds postestimation##estat:estat config}}coordinates of the
	approximating configuration{p_end}
{synopt:{helpb mds postestimation##estat:estat correlations}}correlations
	between dissimilarities and approximating distances{p_end}
{synopt:{helpb mds postestimation##estat:estat pairwise}}pairwise
	dissimilarities, approximating distances, and raw residuals{p_end}
{synopt:{helpb mds postestimation##estat:estat quantiles}}quantiles of the
	residuals per object{p_end}
{synopt:{helpb mds postestimation##estat:estat stress}}Kruskal stress (loss)
	measure{p_end}
{p2coldent:+ {helpb mds postestimation##estat:estat summarize}}estimation
	sample summary{p_end}
{synopt:{helpb mds postestimation##mdsconfig:mdsconfig}}plot of approximating
	configuration{p_end}
{synopt:{helpb mds postestimation##mdsshepard:mdsshepard}}Shepard diagram{p_end}
{synopt:{helpb screeplot}}plot eigenvalues{p_end}
{synoptline}
{p 4 6 2}
+ {cmd:estat summarize} is not available after {cmd:mdsmat}.

{pstd}
In addition, the following standard postestimation commands are available:

{p2coldent:command}description{p_end}
{synoptline}
{p2coldent:* {helpb estimates}}cataloging estimation results{p_end}
{synopt:{helpb mds postestimation##predict:predict}}approximating
	configuration, dissimilarities, distances, and residuals{p_end}
{synoptline}
{p2colreset}{...}
{p 4 6 2}
* All {cmd:estimates} subcommands except {opt table} and {opt stats} are
available.
{p_end}


{title:Special-interest postestimation commands}

{pstd}{cmd:estat config}
lists the coordinates of the approximating configuration.

{pstd}{cmd:estat correlations}
lists the Pearson and Spearman correlations between the dissimilarities and
the approximating distances for each object.

{pstd}{cmd:estat pairwise}
lists the pairwise statistics: the dissimilarities, the approximating
distances, and the raw residuals.

{pstd}{cmd:estat quantiles}
lists the quantiles of the residuals per object.

{pstd}{cmd:estat stress}
displays the Kruskal stress (loss) measure between the (transformed)
dissimilarities and fitted distances per object.

{pstd}{cmd:estat summarize}
summarizes the variables in the MDS over the estimation sample.  After
{helpb mds}, {cmd:estat summarize} also reports whether and how variables were
transformed before computing similarities or dissimilarities.

{pstd}{cmd:mdsconfig}
produces a plot of the approximating Euclidean configuration.  By default,
dimensions 1 and 2 are plotted.

{pstd}{cmd:mdsshepard}
produces a Shepard diagram of the dissimilarities against the approximating
Euclidean distances.  Ideally, the points in the plot should be close to the
"x=y line".  Optionally, separate plots are generated for each "row" (value of
{cmd:id()}).


{marker predict}{...}
{title:Syntax for predict}

{p 8 16 2}
{cmd:predict} {dtype} {newvar:list} {ifin}
[{cmd:,} {it:statistic} {it:options}]

{synoptset 20 tabbed}{...}
{synopthdr:statistic}
{synoptline}
{syntab:Main}
{synopt:{opt con:fig}}approximating configuration; specify {cmd:dimension()} or
	fewer variables{p_end}
{synopt:{opt pair:wise}}dissimilarity, distance, and raw residuals; specify 3
	variables{p_end}
{synoptline}

{synopthdr}
{synoptline}
{syntab:Main}
{p2coldent:* {opt sav:ing(filename)}}save results to {it:filename}{p_end}
{synopt:{opt replace}}overwrite {it:filename}{p_end}
{synopt:{opt notrans:form}}suppress linear transformation of the
	dissimilarities; {opt pairwise} only{p_end}
{synopt:{opt full}}generate all pairs; default is only the ordered pairs,
	excluding the diagonal; {opt pairwise} only{p_end}
{synoptline}
{p2colreset}{...}
{p 4 6 2}
* {opt saving()} is required after {cmd:mdsmat}, after {cmd:mds} if
{opt pairwise} is selected, and after {cmd:mdslong} if {opt config} is
selected.
{p_end}


{title:Options for predict}

{dlgtab:Main}

{phang}{cmd:config}
generates variables containing the approximating configuration in Euclidean
space.  Specify as many new variables as approximating dimensions (as
determined by the {cmd:dimension()} option of {cmd:mds}, {cmd:mdsmat}, or
{cmd:mdslong}), though it is allowed to specify fewer.  Note that
{cmd:estat config} displays the same information, but does not store the
information in variables.  After {cmd:mdsmat} and {cmd:mdslong} you must also
specify the {cmd:saving()} option.

{phang}{cmd:pairwise}
generates three new variables containing pairwise comparison statistics:
dissimilarities, approximating Euclidean distances, and raw residuals.
Specify three new variable names.  Note that {cmd:estat pairwise} displays the
same information, but does not store the information in variables.

{pmore}
After {cmd:mds} and {cmd:mdsmat} you must also specify the {cmd:saving()}
option.  With n objects, the pairwise dataset has n(n-1)/2 observations.
In addition to the three requested variables, {cmd:predict} produces
variables {it:id}{cmd:1} and {it:id}{cmd:2} that identify pairs of objects.
With {cmd:mds}, {it:id} is the name of the identification variable ({cmd:id()}
option) and with {cmd:mdsmat} it is "{cmd:Category}".

{phang}{opt saving(filename)}
is required after {cmd:mdsmat}, after {cmd:mds} if {opt pairwise} is selected,
and after {cmd:mdslong} if {opt config} is selected.  {opt saving()} indicates
that the generated variables are to be created in a new Stata dataset and
saved in the file named {it:filename}.  Unless {opt saving()} is specified,
the variables are generated in the current dataset.  In many cases
{opt saving()} is required; see {opt config} and {opt pairwise} for details.

{phang}{opt replace}
indicates that {it:filename} specified in {cmd:saving()} may be overwritten.

{phang}{opt notransform}
suppresses the linear transformation of the dissimilarities to match the
fitted Euclidean distances.  {cmd:notransform} may only be specified with
{opt pairwise}.

{phang}{opt full}
creates predictions for all pairs of objects (j1,j2).  The default is to
generate predictions only for pairs (j1,j2) with j1>j2.  {opt full} may only
be specified with {opt pairwise}.


{marker estat}{...}
{title:Syntax for estat}

{pstd}
List the coordinates of the approximating configuration

{p 8 14 2}
{cmd:estat} {cmdab:con:fig} [{cmd:,}
{opt max:length(#)} {opth for:mat(%fmt)}]

{pstd}
List the Pearson and Spearman correlations

{p 8 14 2}
{cmd:estat} {cmdab:cor:relations} [{cmd:,}
{opt max:length(#)} {opth for:mat(%fmt)} {opt notot:al}]

{pstd}
List the pairwise statistics

{p 8 14 2}
{cmd:estat} {cmdab:pair:wise} [{cmd:,}
{opt max:length(#)} {opt notrans:form} {opt f:ull} {opt s:eparator}]

{pstd}
List the quantiles of the residuals

{p 8 14 2}
{cmd:estat} {cmdab:qua:ntiles} [{cmd:,}
{opt max:length(#)} {opth for:mat(%fmt)} {opt notot:al} {opt notrans:form}]

{pstd}
Display the Kruskal stress (loss) measure

{p 8 14 2}
{cmd:estat} {cmdab:str:ess} [,
{opt max:length(#)} {opth for:mat(%fmt)} {opt notot:al} {opt notrans:form}]

{pstd}
Summarize the variables in MDS

{p 8 14 2}
{cmd:estat} {cmdab:su:mmarize} [{cmd:,} {opt lab:els}]

{synoptset 16 tabbed}{...}
{synopthdr}
{synoptline}
{syntab:Main}
{synopt:{opt max:length(#)}}maximum number of characters for displaying object
	names; default is {cmd:12}{p_end}
{synopt:{opth for:mat(%fmt)}}display format{p_end}
{synopt:{opt notot:al}}suppress display of overall summary statistics{p_end}
{synopt:{opt notrans:form}}suppress linear transform of the
	dissimilarities{p_end}
{synopt:{opt f:ull}}display all pairs (j1,j2); default is (j1>j2) only{p_end}
{synopt:{opt s:eparator}}draw separating lines{p_end}
{synopt:{opt lab:els}}display variable labels{p_end}
{synoptline}
{p2colreset}{...}


{title:Options for estat}

{dlgtab:Main}