Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
st: Re: st: Some problems concerning nlsur！
From
Jorge Eduardo Pérez Pérez <[email protected]>
To
"[email protected]" <[email protected]>
Subject
st: Re: st: Some problems concerning nlsur！
Date
Thu, 11 Jul 2013 10:56:30 -0400
Some possible answers to some of your problems:
1) and 3) It seems that you are trying to estimate a model with 19
parameters with 20 observations. Some of the parameters may be
completely determined by the initial values and therefore Stata will
not report the standard errors. Or else, since there are few degrees
of freedom. the standard errors of some parameters are too large and
Stata does not report the standard errors.
2) I believe the standard errors you get under the default options are
information matrix standard errors, and those are asymptotic. In any
case, with 20 observations the are not likely to be correct.
_______________________
Jorge Eduardo Pérez Pérez
On Thu, Jul 11, 2013 at 3:40 AM, cham xue <[email protected]> wrote:
> Dear all:
> I'm a student doing my research concerning nlsur，there are several
> doubts bear in my mind from the result that stata gives ：
> FGNLS regression
> ---------------------------------------------------------------------
> Equation | Obs Parms RMSE R-sq Constant
> ----------------+----------------------------------------------------
> 1 lnca | 20 . .0142847 1.0000* (none)
> 2 sl | 20 . .0067662 0.9995* (none)
> 3 se | 20 . .0090048 0.9995* (none)
> ---------------------------------------------------------------------
> * Uncentered R-sq
>
> ------------------------------------------------------------------------------
> | Coef. Std. Err. z P>|z| [95% Conf.
> Interval]
> -------------+----------------------------------------------------------------
> /z0 | -782.0965 . . . .
> .
> /z1 | -.0088672 .0026679 -3.32 0.001 -.0140962
> -.0036382
> /z2 | 5.132203 1.000947 5.13 0.000 3.170383
> 7.094023
> /z14 | .0007173 .0001176 6.10 0.000 .0004869
> .0009478
> /z24 | -.1857547 .049265 -3.77 0.000 -.2823122
> -.0891971
> /z15 | -.0002366 .0000209 -11.34 0.000 -.0002775
> -.0001957
> /z25 | .0545494 .0107154 5.09 0.000 .0335476
> .0755512
> /z11 | .0011035 .0001226 9.00 0.000 .0008631
> .0013438
> /z12 | -.0003213 .000093 -3.46 0.001 -.0005036
> -.0001391
> /z22 | .3592626 .0348429 10.31 0.000 .2909717
> .4275535
> /k1 | .109157 . . . .
> .
> /k2 | 22.75873 . . . .
> .
> /k3 | 35.67095 . . . .
> .
> /b | 1.002294 . . . .
> .
> /z4 | 88.15114 .1810855 486.79 0.000 87.79622
> 88.50606
> /z5 | -13.48022 .1781426 -75.67 0.000 -13.82937
> -13.13107
> /z45 | .7517334 .0089341 84.14 0.000 .7342229
> .769244
> /z44 | -4.868163 .0189082 -257.46 0.000 -4.905223
> -4.831104
> /z55 | -.1081436 .003327 -32.51 0.000 -.1146643
> -.1016228
> ------------------------------------------------------------------------------
>
> Part of the program is presented as follows:
> tempname z21 z22 z23
> tempname z31 z32 z33
> scalar `z11'=`at'[1,8]
> scalar `z12'=`at'[1,9]
> scalar `z13'=-`z11'-`z12'
> scalar `z21' =`z12'
> scalar `z22'=`at'[1,10]
> scalar `z23'=-`z21'-`z22'
> scalar `z31' =`z13'
> scalar `z32'=`z23'
> scalar `z33'=-`z31'-`z32'
> tempname k1 k2 k3
> scalar `k1'=`at'[1,11]
> scalar `k2'=`at'[1,12]
> scalar `k3'=`at'[1,13]
> tempname b z4 z5 z45 z44 z55
> scalar `b'=`at'[1,14]
> scalar `z4'=`at'[1,15]
> scalar `z5'=`at'[1,16]
> scalar `z45'=`at'[1,17]
> scalar `z44'=`at'[1,18]
> scalar `z55'=`at'[1,19]
> quietly{
> replace
> `lnca'=`z0'+`z4'*ln(y)+0.5*`z44'*(ln(y))^2+`z14'*ln(y)*ln(`k1'*pl)+`z24'*ln(y)*ln(`k2'*pk)+`z34'*ln(y)*ln(`k3'*pe)+`z1'*ln(`k1'*pl)+`z2'*ln(`k2'*pk)+`z3'*ln(`k3'*pe)+0.5*`z11'*(ln(`k1'*pl))^2+0.5*`z22'*(ln(`k2'*pk))^2+0.5*`z33'*(ln(`k3'*pe))^2+`z12'*ln(`k1'*pl)*ln(`k2'*pk)+`z13'*ln(`k1'*pl)*ln(`k3'*pe)+`z23'*ln(`k2'*pk)*ln(`k3'*pe)+`z15'*ln(`k1'*pl)*`t'+`z25'*ln(`k2'*pk)*`t'+`z35'*ln(`k3'*`pe')*`t'+`z5'*`t'+0.5*`z55'*`t'^2+`z45'*ln(`y')*`t'+ln((`z1'+`z15'*`t'+`z11'*ln(`k1'*`pl')+`z12'*ln(`k2'*`pk')+`z13'*ln(`k3'*`pe')+`z14'*ln(`y'))/`k1'+(`z2'+`z25'*`t'+`z12'*ln(`k1'*`pl')+`z22'*ln(`k2'*`pk')+`z23'*ln(`k3'*`pe')+`z24'*ln(`y'))/`k2'+(`z3'+`z35'*`t'+`z13'*ln(`k1'*`pl')+`z32'*ln(`k2'*`pk')+`z33'*ln(`k3'*`pe')+`z34'*ln(`y'))/`k3')+ln(1/`b')
> `if'
> replace
> `sl'=((`z1'+`z15'*`t'+`z11'*ln(`k1'*`pl')+`z12'*ln(`k2'*`pk')+`z13'*ln(`k3'*`pe')+`z14'*ln(`y'))/`k1')/((`z1'+`z15'*`t'+`z11'*ln(`k1'*`pl')+`z12'*ln(`k2'*`pk')+`z13'*ln(`k3'*`pe')+`z14'*ln(`y'))/`k1'+(`z2'+`z25'*`t'+`z12'*ln(`k1'*`pl')+`z22'*ln(`k2'*`pk')+`z23'*ln(`k3'*`pe')+`z24'*ln(`y'))/`k2'+(`z3'+`z35'*`t'+`z31'*ln(`k1'*`pl')+`z32'*ln(`k2'*`pk')+`z33'*ln(`k3'*`pe')+`z34'*ln(`y'))/`k3')
> `if'
> replace
> `se'=((`z3'+`z35'*`t'+`z31'*ln(`k1'*`pl')+`z32'*ln(`k2'*`pk')+`z33'*ln(`k3'*`pe')+`z34'*ln(`y'))/`k3')/((`z1'+`z15'*`t'+`z11'*ln(`k1'*`pl')+`z12'*ln(`k2'*`pk')+`z13'*ln(`k3'*`pe')+`z14'*ln(`y'))/`k1'+(`z2'+`z25'*`t'+`z12'*ln(`k1'*`pl')+`z22'*ln(`k2'*`pk')+`z23'*ln(`k3'*`pe')+`z24'*ln(`y'))/`k2'+(`z3'+`z35'*`t'+`z31'*ln(`k1'*`pl')+`z32'*ln(`k2'*`pk')+`z33'*ln(`k3'*`pe')+`z34'*ln(`y'))/`k3')
> `if'
> }
> end
>
> the problems are:
> 1) Why some information of the results of z0、k1、k2、k3、b are missing?
> 2）How to make the stata return asymptotic t-statistics?
> 3) The results are varied from different initial values，why the examples
> from the stata manual didn't give an initial value for the parameters，but
> when I remove the initial values,stata will report an error.
> 4) How to normalize k1 to 1 compared to k2 and k3 before estimation in
> stata?
> 5) Is there any other problem from the result?
>
> Thank you very much for your attention!
> cham xue
>
>
>
> --
> View this message in context: http://statalist.1588530.n2.nabble.com/Some-problems-concerning-nlsur-tp7580394.html
> Sent from the Statalist mailing list archive at Nabble.com.
>
> *
> * For searches and help try:
> * http://www.stata.com/help.cgi?search
> * http://www.stata.com/support/faqs/resources/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/faqs/resources/statalist-faq/
* http://www.ats.ucla.edu/stat/stata/