Re: st: RE: plotting coefficients after quantile regression- beyond grqreg

[Prabhat <prabhat.barnwal@gmail.com>]

Dear STATAlist and Martin,

I think my prob requires separate scripting inside the loop. None of the
options  seem to be helpful here. I am trying for last few hours but still
clueless :(.

Any help will be appreciated.

Thanks.

Regards,
Prabhat




On Mon, Mar 8, 2010 at 8:30 PM, Prabhat <prabhat@iuj.ac.jp> wrote:

> Thank you very much Martin.
> It is very useful. However, I still need one more help. The average I wan=
t
> to use in the equation to predict the dependent variable (i.e. pred 10,
> pred20...etc in this script), should be for those observations belonging =
to
> those quantile only.
>
> I mean, in the place of 'lengthmean', 'lengthmean 10', "lengthmean
> 20"...etc are required for the corresponding quantile of dependent variab=
le
> i.e. price here.
>
> I tried to move egen inside the for loop, but I failed to figure out the
> right way.
>
> Could you suggest a way out?
>
> Thanks again.
>
> Regards,
> Prabhat
> International University of Japan
>
>
>
> On Mon, Mar 8, 2010 at 3:10 AM, Martin Weiss <martin.weiss1@gmx.de> wrote=
:
>
>>
>> <>
>>
>> The difficulty here is that -qreg- does not seem to like -fvvarlist-s,
>> which
>> in other cases make quadratic vars much easier to specify (It does not
>> allow
>> -coeflegend-, either...). So you want to check this very, very carefully
>> whether it does what you want:
>>
>> *******
>> sysuse auto, clear
>> gen mpgsq=3Dmpg^2
>> egen lengthmean=3Dmean(length)
>> egen turnmean=3Dmean(turn)
>>
>> foreach quant of numlist 10 25 50 75 90{
>>        qreg price mpg mpgsq length turn, q(`quant')
>>        gen pred`quant'=3D_b[_cons]+_b[mpg]*mpg+_b[mpgsq]*mpgsq+ /*
>>         */ _b[length]*lengthmean+_b[turn]*turnmean
>>        la var pred`quant' "Quantile `quant'"
>> }
>>
>> sort mpg
>>
>> tw (line pred10 mpg)  /*
>>  */ (line pred25 mpg) (line pred50 mpg)  /*
>>  */ (line pred75 mpg) (line pred90 mpg)
>> *******
>>
>>
>> HTH
>> Martin
>>
>>
>> -----Original Message-----
>> From: owner-statalist@hsphsun2.harvard.edu
>> [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Prabhat
>> Sent: Sonntag, 7. M=E4rz 2010 15:04
>> To: statalist@hsphsun2.harvard.edu
>> Subject: st: plotting coefficients after quantile regression- beyond
>> grqreg
>>
>> Dear all,
>>
>> I need your help regarding plots for quantile regression  results. I am
>> new
>> to STATA as well as quantile regression both.
>>
>> I am running a quantile regression with independent variable X and its
>> square term i.e. X^2. there are a number of other variables. however, my
>> requirement is to show the parabolic curve i.e. dependent variable vs X
>> for
>> quantile=3D 0.1,0.25, 0.5, 0.75, 0.9,  keeping all other explanatory
>> variables
>> at their average value. So, it will be kind of a 5 parabolic curves on a
>> single plot.
>>
>> I know there is a very helpful command GRQREG, but it does not plot the
>> particular estimated coefficients for a number of quantiles on the same
>> plot. Kindly correct me, if there is any way here.
>>
>> Further, the only other option I can think of if to do it manually after
>> getting the results. But, that seems to be a  difficult and tedious task
>> because i have to repeat it for a number of cycles.
>>
>> Any suggestion will be highly appreciated.
>>
>> Thank you.
>>
>> Regards.
>> Prabhat Barnwal
>> International University of Japan
>>
>> *
>> *   For searches and help try:
>> *   http://www.stata.com/help.cgi?search
>> *   http://www.stata.com/support/statalist/faq
>> *   http://www.ats.ucla.edu/stat/stata/
>>
>>
>> *
>> *   For searches and help try:
>> *   http://www.stata.com/help.cgi?search
>> *   http://www.stata.com/support/statalist/faq
>> *   http://www.ats.ucla.edu/stat/stata/
>>
>
>

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/