procrustes.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}
{* 06apr2005}{...}
{cmd:help procrustes} {...}
{right:dialog:  {bf:{dialog procrustes}}{space 15}}
{right:also see:  {help procrustes postestimation}}
{hline}

{title:Title}

{p 4 25 2}
{hi:[MV] procrustes} {hline 2} Procrustes transformation


{title:Syntax}

{p 8 24 2}{cmd:procrustes}
{cmd:(}{it:{help varlist:varlist_y}}{cmd:)}
{cmd:(}{it:{help varlist:varlist_x}}{cmd:)}
{ifin} {weight} [{cmd:,} {it:options} ]

{synoptset 27 tabbed}{...}
{synopthdr}
{synoptline}
{syntab:Model}
{synopt:{cmdab:tr:ansform:(}{cmdab:or:thogonal}{cmd:)}}orthogonal rotation and
	reflection transformation; the default{p_end}
{synopt:{cmdab:tr:ansform:(}{cmdab:ob:lique}{cmd:)}}oblique rotation
	transformation{p_end}
{synopt:{cmdab:tr:ansform:(}{cmdab:un:restricted}{cmd:)}}unrestricted
	transformation{p_end}
{synopt:{opt nocon:stant}}suppress the constants{p_end}
{synopt:{opt norh:o}}suppress the dilation factor, i.e., set rho=1{p_end}
{synopt:{opt force}}allow overlap and duplicates in {it:varlist_y}
	and {it:varlist_x} (advanced){p_end}

{syntab:Reporting}
{synopt:{opt nofi:t}}suppress table of fit statistics by target variable{p_end}
{synoptline}
{p2colreset}{...}

{pstd}
{cmd:bootstrap}, {cmd:by}, {cmd:jackknife}, and {cmd:statsby} are
allowed; see {help prefix}.
{p_end}
{pstd}
{cmd:aweight}s and {cmd:fweight}s are allowed; see {help weights}.
{p_end}
{pstd}
See {help procrustes postestimation} for features available after estimation.
{p_end}


{title:Description}

{pstd}
{cmd:procrustes} performs the Procrustean analysis, one of the standard
methods of multidimensional scaling.  For given "target" variables
{it:varlist_y} and "source" variables {it:varlist_x}, the goal is to transform
the source {bf:X} to be as close as possible to the target {bf:Y}.  The
permitted transformations are any combination of dilation (uniform scaling),
rotation and reflection (i.e., orthogonal or oblique transformations), and
translation.  Closeness is measured by the residual sum of squares.

{pstd}
Formally, {cmd:procrustes} solves the minimization problem

{pin2}
Minimize {space 1}  | {bf:Y} - ({bf:1} {bf:c}' + rho {bf:X} {bf:A}) |

{pstd}
where {bf:c} is a row-vector representing the translation; rho is the
scalar "dilation factor"; {bf:A} is the rotation and reflection matrix
(orthogonal, oblique, or unrestricted), and |.| denotes the L2 norm.

{pstd}
The number of source ({bf:X}) variables should not exceed the number of target
({bf:Y}) variables for the orthogonal transformation.  {cmd:procrustes} deals
with complete cases only.  Efficient algorithms for the "partial Procrustes
problem" in which some elements of the target variables are missing are still
under development.

{pstd}
{cmd:procrustes} assumes equal weights or scaling for the dimensions.  It
would be inappropriate, for example, to have the first variable measured in
grams ranging from 5000 to 8000, the second variable measured in dollars
ranging from 3 to 12, and the third variable measured in meters ranging from
100 to 280.  In such cases you would want to operate on standardized
variables.


{title:Options}

{dlgtab:Model}

{phang}{opt transform(transform)}
specifies the transformation method.  The following transformation methods
are allowed:

{phang2}{opt orthogonal}
specifies that the linear transformation matrix {bf:A} should be orthogonal,
{bf:A}'{bf:A} = {bf:A}{bf:A}' = {bf:I}.  This is the default.

{phang2}{opt oblique}
specifies that the linear transformation matrix {bf:A} should be oblique,
diag({bf:A}{bf:A}') = {bf:1}.

{phang2}{opt unrestricted}
applies no restrictions to {bf:A}.  The procrustes transformation is now
equivalent to multivariate regression with uncorrelated errors; see
{helpb mvreg}.

{phang}{opt noconstant}
specifies that the translation component {bf:c} is fixed at {bf:0}
(the 0-vector).

{phang}{opt norho}
specifies that the dilation (scaling) constant rho is fixed at 1.
This option is not relevant with {cmd:transform(unrestricted)}; in
this case rho is always fixed at 1.

{phang}{opt force},
an advanced option,
allows overlap and duplicates in the target variables {it:varlist_y}
and source variables {it:varlist_x}.

{dlgtab:Reporting}

{phang}{opt nofit}
suppresses the table with fit statistics per target variable.  This option may
be specified during estimation or replay.


{title:Examples}

    {cmd:. procrustes (y1 y2 y3) (x1 x2 x3)}

    {cmd:. procrustes (y*) (x1 x2) if z>0, transform(oblique)}

    {cmd:. procrustes (y1-y3) (x1 x2 x3), norho}


{title:Also see}

{psee}
Manual:  {bf:[MV] procrustes}

{psee}
Online:  {help procrustes postestimation};{break}
{helpb ca},
{helpb mvreg},
{helpb pca},
{helpb rotate}
{p_end}