mlogit_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}
{* *! version 1.0.0  10jun2005}{...}
{cmd:help mlogit postestimation}{right:dialog:  {bf:{dialog mlogit_p:predict}}}
{right:also see:  {helpb mlogit} }
{hline}

{title:Title}

{p2colset 5 34 36 2}{...}
{p2col :{hi:[R] mlogit postestimation} {hline 2}}Postestimation tools for mlogit{p_end}
{p2colreset}{...}


{title:Description}

{pstd}
The following postestimation commands are available for {opt mlogit}:

{synoptset 11}{...}
{p2coldent :command}description{p_end}
{synoptline}
INCLUDE help post_estat
INCLUDE help post_estimates
INCLUDE help post_hausman
INCLUDE help post_lincom
INCLUDE help post_lrtest
INCLUDE help post_mfx
INCLUDE help post_nlcom
{synopt :{helpb mlogit postestimation##predict:predict}}predictions, residuals,
influence statistics, and other diagnostic measures{p_end}
INCLUDE help post_predictnl
INCLUDE help post_suest
INCLUDE help post_test
INCLUDE help post_testnl
{synoptline}
{p2colreset}{...}


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

{p 8 16 2}
{cmd:predict}
{dtype}
{newvar}
{ifin}
[{cmd:,}
{it:statistic}
{opt o:utcome(outcome)}]

{p 8 16 2}
{cmd:predict}
{dtype}
{c -(}{it:stub*}{c |}{it:newvar_1} ... {it:newvar_k-1}{c )-}
{ifin}
{cmd:,}
{opt sc:ores}

{phang}
where k is the number of outcomes in the model.

{synoptset 11 tabbed}{...}
{synopthdr :statistic}
{synoptline}
{syntab :Main}
{synopt :{opt p:r}}probability of a positive outcome; the default{p_end}
{synopt :{cmd:xb}}xb, fitted values{p_end}
{synopt :{cmd:stdp}}standard error of the linear prediction{p_end}
{synopt :{cmd:stddp}}standard error of the difference in two linear
predictions{p_end}
{synoptline}
{p2colreset}{...}
{p 4 6 2}
Note that you specify one new variable with {opt xb}, {opt stdp}, and
{opt stddp} and specify either one or k new variables with {opt p}.
{p_end}
INCLUDE help esample


{title:Options for predict}

{phang}
{opt pr}, the default, calculates the probability of a positive outcome
conditional on one positive outcome within group.

{pmore}
If you do not also specify the {opt outcome(outcome)} option, you must
    specify k new variables.  For instance, say that you fitted your model by
    typing {cmd:mlogit insure age male} and that {opt insure} takes on three
    values.  Then you could type {cmd:predict p1 p2 p3, pr} to obtain all
    three predicted probabilities.  It does not matter which category
    {opt mlogit} chooses as the base outcome; {opt predict} will
    calculate all three probabilities correctly.

{pmore}
If you also specify the {opt outcome(outcome)} option, then you specify one
    new variable.  Say that {opt insure} took on values 1, 2, and 3.
    Typing {cmd:predict p1, pr outcome(1)} would produce the same {opt p1} as
    above, {cmd:predict p2, pr outcome(2)} the same {opt p2} as above; and so
    on.  If {opt insure} took on values 7, 22, and 93, you would specify
    {cmd:outcome(7)}, {cmd:outcome(22)}, and {cmd:outcome(93)}.

{phang}
{opt xb} calculates the linear prediction.  You must also specify the
{opt outcome(outcome)} option.

{phang}
{opt stdp} calculates the standard error of the linear prediction.
You must also specify the {opt outcome(outcome)} option.

{phang}
{opt stddp} calculates the standard error of the difference in two
linear predictions.  You must specify option {opt outcome(outcome)},
and in this case, you specify the two particular outcomes of interest inside
the parentheses, for example, {cmd:predict sed, stdp outcome(1,3)}.

{phang}
{opt outcome(outcome)} specifies the outcome for which the statistic is to be
calculated.  {opt equation()} is a synonym for {opt outcome()}: it does not
matter which you use, and the standard rules for specifying an {opt equation()}
apply.

{phang}
{opt scores} calculates equation-level score variables.  The number of
score variables created will be one less than the number of outcomes in the
model.  If the number of outcomes in the model were k, then

{pmore}
the first new variable will contain the first derivative of the log
likelihood with respect to the first equation;

{pmore}
the second new variable will contain the first derivative of the log
likelihood with respect to the second equation;

{pmore}
...

{pmore}
the (k-1)st new variable will contain the first derivative of the log
likelihood with respect to the (k-1)st equation.


{title:Marginal effects}

{pstd}
{cmd:predict} after {cmd:mlogit}, unlike most other estimation commands,
can predict multiple new variables
by issuing {cmd:predict} only once.  To calculate
marginal effects for the probability
of more than one outcome, run {cmd:mfx} separately for
each outcome, as shown in the example below.


{title:Examples}

{phang}{cmd:. webuse sysdsn3, clear}{p_end}
{phang}{cmd:. mlogit insure age male nonwhite site2 site3}{p_end}
{phang}{cmd:. predict p1 if e(sample), outcome(1)}{p_end}
{phang}{cmd:. summarize p1}{p_end}
{phang}{cmd:. predict idx1, outcome(Indemnity) xb}{p_end}

{phang}{cmd:. sysuse auto, clear}{p_end}
{phang}{cmd:. mlogit rep78 mpg displ, nolog}{p_end}
{phang}{cmd:. mfx, predict(p outcome(1))}{p_end}
{phang}{cmd:. mfx, predict(p outcome(2))}{p_end}
{phang}{cmd:. mfx, predict(p outcome(3))}{p_end}

{title:Also see}

{psee}
Manual:  {bf:[R] mlogit}

{psee}
Online:  {helpb mlogit};{break}
{helpb estimates},
{helpb hausman},
{helpb lincom},
{helpb lrtest},
{helpb mfx},
{helpb nlcom},
{helpb predictnl},
{helpb suest},
{helpb test},
{helpb testnl}
{p_end}