_predict.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}
{* 09feb2005}{...}
{cmd:help _predict}
{hline}

{title:Title}

{p2colset 5 21 23 2}{...}
{p2col :{hi:[P] _predict} {hline 2}}Obtain predictions, residuals, etc., after estimation programming command{p_end}
{p2colreset}{...}


{title:Syntax}

{phang}After {cmd:regress}

{p 8 17 2}{cmd:_predict} {dtype} {newvar}
        {ifin} [{cmd:,} {opt xb} {opt stdp} {opt stdf} {opt stdr}
        {opt h:at} {opt c:ooksd} {opt re:siduals} {opt rsta:ndard}
        {opt rstu:dent} {opt nooff:set} {opt nolab:el}]

{phang}After single-equation (SE) estimators

{p 8 17 2}{cmd:_predict} {dtype} {newvar}
        {ifin} [{cmd:,} {opt xb} {opt stdp} {opt nooff:set} {opt nolab:el}]

{phang}After multiple-equation (ME) estimators

{p 8 17 2}{cmd:_predict} {dtype} {newvar}
        {ifin} [{cmd:,} {opt xb} {opt stdp} {opt stddp}
        {opt nooff:set} {opt nolab:el} 
        {cmdab:e:quation(}[{cmd:,} {it:eqno}]{cmd:)}]

{pstd}
If no prediction options are specified, the following are calculated.

{p 14 32 2}after {cmd:probit} {space 9} probability of a positive outcome{p_end}
{p 14 32 2}after {cmd:logit} {space 10} probability of a positive outcome{p_end}
{p 14 32 2}after {cmd:mlogit} {space 9} probability of a positive outcome{p_end}
{p 14 32 2}after {cmd:stcox} {space 10} relative hazard{p_end}
{p 14 32 2}after others commands{space 2}linear prediction


{title:Description}

{pstd}
{cmd:_predict} is for use by programmers as a subroutine for implementing
the {cmd:predict} command for use after estimation; see {helpb predict}.


{title:Options}

{phang}
{cmd:xb} calculates the linear prediction from the estimated model.  That is,
all models can be thought of as estimating a set of parameters b1, b2, ..., bk,
and the linear prediction is y = xb.  In the case of linear regression, the
values are called the predicted values, or, for, out-of-sample predictions, 
the forecast.  In the case of logit and probit, for example, y is called the 
logit or probit index.

{pmore}
It is important to understand that the x1, x2, ..., xk used in the calculation
are obtained from the data currently in memory and do not have to correspond
to the data on the independent variables used in estimating the model
(obtaining the b1, b2, ..., bk).

{phang}
{cmd:stdp} calculates the standard error of the prediction after any
estimation command.  Here the prediction is understood to mean the same thing
as the "index", namely, xb.  The statistic produced by {cmd:stdp} can be
thought of as the standard error of the predicted expected value, or mean
index, for the observation's covariate pattern.  This is also commonly
referred to as the standard error of the fitted value.

{phang}
{cmd:stdf} calculates the standard error of the forecast, which is the standard error of the point prediction for a single observation.  It is commonly 
referred to as the standard error of the future or forecast value.
By construction, the standard errors produced by {cmd:stdf} are always larger
than those by {cmd:stdp}.

{phang} {cmd:stdr} calculates the standard error of the residuals.

{phang}
{cmd:stdr} calculates the standard error of the residuals.

{phang}
{cmd:stddp} is allowed only after you have previously estimated a
multiple-equation model.  The standard error of the difference in linear
predictions between equations 1 and 2 is calculated.

{phang}
{cmd:hat} or {cmd:leverage} calculates the diagonal elements of the
projection hat matrix.

{phang}
{cmd:cooksd} calculates Cook's D influence statistic.

{phang}
{cmd:residuals} calculates the residuals.

{phang}
{cmd:rstandard} calculates the standardized residuals.

{phang}
{cmd:rstudent} calculates the studentized (jackknifed) residuals.

{phang}
{cmd:nooffset} may be combined with most statistics and specifies that the
calculation be made, ignoring any offset or exposure variable specified when
the model was estimated.

{pmore}
This option is available, even if not documented, for {cmd:predict} after a
specific command.  If neither the {opt offset(varname)} option nor the 
{opt exposure(varname)} option was specified when the model was estimated,
specifying {opt nooffset} does nothing.

{phang}
{cmd:nolabel} prevents {cmd:_predict} from labeling the newly created
variable.

{phang}
{cmd:equation(}{it:eqno}[{cmd:,}{it:eqno}]{cmd:)} is relevant only when you
have previously estimated a multiple-equation model.  It specifies the
equation to which you are referring.

{pmore}
{cmd:equation()} is typically filled in with one {it:eqno}{hline 2}it would be
filled in that way with options {cmd:xb} and {cmd:stdp}, for instance.  
{cmd:equation(#1)} would mean that the calculation is to be made for the first
equation, {cmd:equation(#2)} would mean the second, and so on.  Alternatively,
you could refer to the equations by their names.  {cmd:equation(income)} would
refer to the equation name {cmd:income} and {cmd:equation(hours)} to the
equation named hours.

{pmore}
If you do not specify {cmd:equation()}, the results are the same as if you
specified {cmd:equation(#1)}.

{pmore}
Other statistics refer to between-equation concepts; {cmd:stddp} is an
example.  In those cases, you might specify {cmd:equation(#1,#2)} or
{cmd:equation(income,hours)}.  When two equations must be specified,
{cmd:equation()} is required.


{title:Also see}

{psee}
Manual:  {bf:[P] _predict}

{psee}
Online:  {helpb predict}, {helpb _pred_se}
{p_end}