help procrustes postestimation dialogs: predict estat
procoverlay
also see: procrustes
-------------------------------------------------------------------------------
Title
[MV] procrustes postestimation -- Postestimation tools for procrustes
Description
The following postestimation commands are of special interest after
procrustes:
command description
-------------------------------------------------------------------------
estat compare fit statistics for orthogonal, oblique, and
unrestricted transformations
estat mvreg display multivariate regression resembling
unrestricted transformation
estat summarize display summary statistics over the estimation
sample
procoverlay produce a Procrustes overlay graph
-------------------------------------------------------------------------
The following standard postestimation commands are also available:
command description
-------------------------------------------------------------------------
* estimates catalog estimation results
predict compute fitted values and residuals
-------------------------------------------------------------------------
* All estimates subcommands except table and stats are available.
Special-interest postestimation commands
estat compare displays a table with fit statistics of the three
transformations provided by procrustes: orthogonal, oblique, and
unrestricted. The two additional procrustes analyses are performed on
the same sample as the original procrustes analysis and with the same
options. F tests comparing the models are provided.
estat mvreg produces the mvreg output related to the unrestricted
Procrustes analysis (the transform(unrestricted) option of procrustes).
estat summarize displays summary statistics over the estimation sample of
the target and source variables (varlist_y and varlist_x).
procoverlay displays a plot of the target variables overlaid with the
fitted values derived from the source variables. If there are more than
two target variables, multiple plots are shown in one graph.
Syntax for predict
predict [type] {stub*|newvarlist} [if] [in] [, statistic ]
statistic description
-------------------------------------------------------------------------
Main
fitted fitted values 1 c' + rho X A; the default (specify #y
vars)
residuals unstandardized residuals (specify #y vars)
q residual sum of squares over the target variables (specify
one var)
-------------------------------------------------------------------------
These statistics are available both in and out of sample; type predict
... if e(sample) ... if wanted only for the estimation sample.
Menu
Statistics > Postestimation > Predictions, residuals, etc.
Options for predict
+------+
----+ Main +-------------------------------------------------------------
fitted, the default, computes fitted values, i.e., the least squares
approximations of the target (varlist_y) variables. You must specify
the same number of new variables as there are target variables.
residuals computes the raw (unstandardized) residuals for each target
(varlist_y) variable. You must specify the same number of new
variables as there are target variables.
q computes the residual sum of squares over all variables, i.e., the
squared Euclidean distance between the target and transformed source
points. Specify one new variable.
Syntax for estat
Table of fit statistics
estat compare [, detail ]
Comparison of mvreg and procrustes output
estat mvreg [, mvreg_options ]
Display summary statistics
estat summarize [, label noheader noweights ]
Menu
Statistics > Postestimation > Reports and statistics
Options for estat
detail, an option with estat compare, displays the standard procrustes
output for the two additional transformations.
mvreg_options, allowed with estat mvreg, are any of the options allowed
by mvreg; see [R] mvreg. The constant is already suppressed if the
Procrustes analysis suppressed it.
label, noheader, and noweights are the same as for the generic estat
summarize command; see [R] estat.
Syntax for procoverlay
procoverlay [if] [in] [, procoverlay_options ]
procoverlay_options description
-------------------------------------------------------------------------
Main
autoaspect adjust aspect ratio on the basis of the
data; default aspect ratio is 1
targetopts(target_opts) affect the rendition of the target
sourceopts(source_opts) affect the rendition of the source
Y axis, X axis, Titles, Legend, Overall
twoway_options any options other than by() documented in
[G] twoway_options
By
byopts(by_option) affect the rendition of combined graphs
-------------------------------------------------------------------------
target_opts description
-------------------------------------------------------------------------
Main
nolabel removes the default observation label from
the target
marker_options change look of markers (color, size, etc.)
marker_label_options change look or position of marker labels
-------------------------------------------------------------------------
source_opts description
-------------------------------------------------------------------------
Main
nolabel removes the default observation label from
the source
marker_options change look of markers (color, size, etc.)
marker_label_options change look or position of marker labels
-------------------------------------------------------------------------
Menu
Statistics > Multivariate analysis > Procrustes overlay graph
Options for procoverlay
+------+
----+ Main +-------------------------------------------------------------
autoaspect specifies that the aspect ratio be automatically adjusted
based on the range of the data to be plotted. This option can make
some procoverlay plots more readable. By default, procoverlay uses
an aspect ratio of one, producing a square plot.
As an alternative to autoaspect, the twoway_option aspectratio() can
be used to override the default aspect ratio. procoverlay accepts
the aspectratio() option as a suggestion only and will override it
when necessary to produce plots with balanced axes, i.e., where
distance on the x axis equals distance on the y axis.
twoway_options, such as xlabel(), xscale(), ylabel(), and yscale(),
should be used with caution. These options are accepted but may have
unintended side effects on the aspect ratio.
targetopts(target_opts) affects the rendition of the target plot. The
following target_opts are allowed:
nolabel removes the default target observation label from the graph.
marker_options affect the rendition of markers drawn at the plotted
points, including their shape, size, color, and outline; see [G]
marker_options.
marker_label_options specify if and how the markers are to be
labeled; see [G] marker_label_options.
sourceopts(source_opts) affects the rendition of the source plot. The
following source_opts are allowed:
nolabel removes the default source observation label from the graph.
marker_options affect the rendition of markers drawn at the plotted
points, including their shape, size, color, and outline; see [G]
marker_options.
marker_label_options specify if and how the markers are to be
labeled; see [G] marker_label_options.
+-----------------------------------------+
----+ Y axis, X axis, Titles, Legend, Overall +--------------------------
twoway_options are any of the options documented in [G] twoway_options,
excluding by(). These include options for titling the graph (see [G]
title_options) and for saving the graph to disk (see [G]
saving_option).
See autoaspect above for a warning against using options such as
xlabel(), xscale(), ylabel(), and yscale().
+----+
----+ By +---------------------------------------------------------------
byopts(by_option) is documented in [G] by_option. This option affects
the appearance of the combined graph and is ignored, unless there are
more than two target variables specified in procrustes.
Examples
Setup
. webuse speed_survey
. procrustes (survey_x survey_y) (speed_x speed_y)
Compare transformations
. estat compare
Produce Procrustes overlay graph
. procoverlay
Saved results
estat compare after procrustes saves the following in r():
Matrices
r(cstat) Procrustes statistics, degrees of freedom, and RMSEs
r(fstat) F statistics, degrees of freedom, and p-values
estat mvreg does not return results.
estat summarize after procrustes saves the following in r():
Matrices
r(stats) means, standard deviations, minimums, and maximums
See also
Manual: [MV] procrustes postestimation
Help: [MV] procrustes;
[R] mvreg
