estcom.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}
{* 07apr2005}{...}
{cmd:help estcom}
{hline}

{title:Title}

{p 4 13 2}
{hi:[U] 20 Estimation and postestimation commands}


{title:Description}

{pstd}For a list of Stata's estimation commands see
{help estimation commands}.  For a discussion of postestimation
commands see {help postest}.

{pstd}
Properties shared by all estimation commands are listed.


{title:Remarks}

{pstd}
Stata commands that fit statistical models{hline 2}commands such as
{helpb logit}, {helpb regress}, {helpb logistic}, 
{helpb sureg}, etc.{hline 2}work similarly.


{title:1. Common syntax}

{pstd}
Single-equation estimation commands usually have syntax,

{p 8 16 2}
{it:command} {varlist} {ifin} {weight} [{cmd:,} {it:options}]

{pstd}
and multiple-equation estimation commands usually have syntax

{p 8 16 2}
{it:command} {cmd:(}{varlist}{cmd:)} {it:...}
{cmd:(}{varlist}{cmd:)} {ifin} {weight}
[{cmd:,} {it:options}]

{pstd}
In single-equation commands, the first variable in {it:varlist} is the
dependent variable and the remaining are the independent variables, although
there can be variations.  For instance, {helpb anova} allows you to specify
terms in addition to variables for the independent variables.


{title:2. Estimation on subsamples}

{pstd}
You can use Stata's standard syntax ({it:{help if}} and {it:{help in}})
to restrict the sample; 
you do not have to make a special dataset.


{title:3. Robust variance estimates}

{pstd}
Most estimation commands allow option {cmd:robust} that provides the
Huber/White/Sandwich estimator of variance.  Those that do also provide option
{cmd:cluster()} that relaxes the assumption of independence.  See 
{hi:[U] 20.14 Obtaining robust variance estimates}.


{title:4. Prefix commands}

{pstd}
Prefix commands may be used to modify the estimation command.  The syntax is

{p 8 16 2}
{it:prefix_command} {it:...} {cmd::} {it:command} {it:...}

{pstd}
where

{p2colset 9 31 33 2}{...}
{p2col:{it:prefix_command}}description{p_end}
{p2line}
{p2col:{helpb bootstrap:bootstrap:}}bootstrap estimation{p_end}
{p2col:{helpb jackknife:jackknife:}}jackknife estimation{p_end}
{p2col:{helpb rolling:rolling:}}collect statistics on moving subsets{p_end}
{p2col:{helpb statsby:statsby:}}collect results across subsets of the data{p_end}
{p2col 7 31 33 2:* {helpb svy:svy:}}estimation with complex survey data{p_end}
{p2col 7 31 33 2:* {helpb stepwise:stepwise:}}stepwise estimation{p_end}
{p2col:{helpb xi:xi:}}interaction expansion{p_end}
{p2line}
{p2col:* Available for most but not all estimation commands.}{p_end}
{p2col:See help {help prefix} for a full list of {it:prefix_commands}.}{p_end}
{p2colreset}{...}

{pstd}
Before using the {cmd:bootstrap:} or {cmd:jackknife:} prefixes, however, 
check whether the estimation command allows option 
{cmd:vce(bootstrap)} or {cmd:vce(jackknife)}.  If it does, it is better 
to use the option rather than the prefix.  The option is implemented in 
terms of the prefix command, but the option automatically knows to pass all
the appropriate suboptions for the specific estimator you are using.


{title:5. Confidence intervals of parameters}

{pstd}
Estimation commands display confidence intervals of
the coefficients.  Estimation-command option {cmd:level()} specifies the width
of the interval.  The default is {cmd:level(95)}, meaning 95% confidence
intervals.

{pstd}
You can reset the default with {helpb level:set level}.


{title:6. Tests of parameters}

{pstd}
You can perform tests on the estimated parameters using 

{p 8 12 2}
o{space 2}{helpb test} {hline 2} 
Wald test of linear hypotheses 

{p 8 12 2}
o{space 2}{helpb testnl} {hline 2} Wald test of nonlinear hypotheses

{p 8 12 2}
o{space 2}{helpb lrtest} {hline 2} likelihood-ratio tests 

{p 8 12 2}
o{space 2}{helpb hausman} {hline 2} Hausman's specification test

{p 8 12 2}
o{space 2}{helpb suest} {hline 2} generalization of the Hausman test


{title:7. Point estimates and CIs of linear and nonlinear combinations}

{pstd}
You can obtain point estimates
and confidence intervals of linear combinations of the estimated parameters
using {helpb lincom}, and of nonlinear combinations using {helpb nlcom}.


{title:8. Predictions}

{pstd}
You can obtain predictions, residuals, influence
statistics, and the like, either for the data on which you just estimated or
for some other data, using {cmd:predict}.

{pstd}
The help for {cmd:predict} is found in two places:

{p 8 12 2}
1.
help {helpb predict} {hline 2} general information

{p 8 12 2}
2.  help {it:estimation_command} {cmd:postestimation} {hline 2} specific 
information and special features following estimation by 
{it:estimation_command}.
For instance, help {helpb regress postestimation}
tells you about {cmd:predict} following {cmd:regress}.

{pstd}
The easy way to access the postestimation help is to see
help {helpb regress} (or whatever estimation command you are using) 
and then select {it:postestimation}.


{title:9. Generalized predictions}

{pstd}
You can obtain nonlinear predictions, 
standard errors, Wald test statistics, significance levels, and confidence
intervals, either for the data on which you just estimated or for some other
data, using {helpb predictnl}.

{pstd}
One especially useful feature of {cmd:predictnl} is that you can obtain
standard errors for most predictions available via {cmd:predict}, and
you can also obtain standard errors of functions and combinations of these
predictions.


{title:10. Marginal effects}

{pstd}
Command {helpb mfx} displays model results in
terms of marginal effects ({it:dy}/{it:dx} or even {it:d}f({it:y})/{it:dx}), which can be displayed
as either derivatives or elasticities.


{title:11. Adjusted means}

{pstd}
Command {helpb adjust} displays tables of adjusted means.


{title:12. Estimation statistics}

{pstd}
Command {helpb estat} displays 
scalar and matrix valued
postestimation statistics such as AIC and BIC.


{title:13. Variance-covariance matrix of the estimators (VCE)}

{pstd}
Command {helpb estat:estat vce} displays the VCE{hline 2}either as a 
covariance matrix or as a correlation matrix.

{pstd}
You can obtain the coefficients and VCE into Stata matrices using {cmd:e(b)}
and {cmd:e(V)} in expressions.

{pstd}
You can obtain the coefficients and VCE into Mata matrices using 
{cmd:st_matrix(e(b))}
and
{cmd:st_matrix(e(V))};
see {helpb mata st_matrix():[M-5] st_matrix()}.


{title:14. Coefficients and standard errors in expressions}

{pstd}
You can refer to the coefficients and standard errors in {help expressions}
using {cmd:_b[}{it:name}{cmd:]} and using {cmd:_se[}{it:name}{cmd:]}, such as

	. {cmd:generate contribution = _b[mpg]*mpg}

{pstd}
See 
{hi:[U] 13.5 Accessing coefficients and standard errors} and 
see {help _b}.


{title:15. Managing and combining estimates}

{pstd}
You can store estimation results using command 
{helpb estimates:estimates store}.
These estimation results may later be restored, replayed, the coefficients of
one or more may be combined in a table, etc.; see {helpb estimates}.

{pstd}
Programmers should also see command {helpb _estimates}, which is a low-level
tool that manages stored estimation results.


{title:16. Redisplaying estimates}

{pstd}
You can, at any time, review your most recent estimates by typing the
estimation command without arguments.


{title:17. Factor rotations}

{pstd}
You can rotate loadings after factor-like commands; 
see {helpb rotate}.


{title:18. Specialized graphs}

{pstd}
There are specialized graph commands available after some estimation commands.

{pstd}
For instance, command {helpb lroc} will graph the ROC curve after {cmd:logit}
or {cmd:probit}.  Command {helpb screeplot} will make scree
plots after {cmd:factor} or {cmd:pca}.

{pstd}
What is available can always be found in the postestimation section 
of the documentation following the estimator.


{title:Also see}

{psee}
Manual:  {bf:[U] 20 Estimation and postestimation commands},{break}
{bf:[U] 26 Overview of Stata estimation commands},{break}
{bf:[I] estimation commands}

{psee}
Online:  {help estimation commands}, {help postest};{break}
{helpb estimates}, {helpb estat};
{helpb adjust}, {helpb bootstrap},
{helpb hausman}, {helpb jackknife}, {help level},
{helpb lincom}, {helpb linktest},
{helpb lrtest}, {helpb mfx}, {helpb nlcom},
{helpb permute}, {helpb predict},
{helpb predictnl}, {helpb rotate},
{helpb simulate}, {helpb statsby},
{helpb suest}, {helpb svy},
{helpb stepwise}, {helpb test},
{helpb testnl}
{p_end}