Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: RE : RE: RE: Problem with "ereturn post" command


From   Valerie Orozco <Valerie.Orozco@toulouse.inra.fr>
To   "statalist@hsphsun2.harvard.edu" <statalist@hsphsun2.harvard.edu>
Subject   st: RE : RE: RE: Problem with "ereturn post" command
Date   Fri, 5 Feb 2010 18:18:55 +0100

Thanks for your response Nick, but it seems that b1 and Sameb1, and then V1 and SameV1 are the same (except if the 0s displayed are not exactly 0) :

. mat diffV = SameV1 - V1

. mat diffb = Sameb1 - b1

. mat list diffV

symmetric diffV[8,8]
         e1vs3   e1vs1  eh1vs1   e2vs3   e2vs2  eh2vs2   e2vs1   e1vs2
 e1vs3       0
 e1vs1       0       0
eh1vs1       0       0       0
 e2vs3       0       0       0       0
 e2vs2       0       0       0       0       0
eh2vs2       0       0       0       0       0       0
 e2vs1       0       0       0       0       0       0       0
 e1vs2       0       0       0       0       0       0       0       0

. mat list diffb

diffb[1,8]
     e1vs3   e1vs1  eh1vs1   e2vs3   e2vs2  eh2vs2   e2vs1   e1vs2
r1       0       0       0       0       0       0       0       0

. scalar d = det(V1)

. scalar samed=det(SameV1)

. mat t = inv(SameV1)

. scalar list
     samed = -2.924e-63
         d = -2.924e-63

. eret post Sameb1 SameV1
Warning:  variance matrix is nonsymmetric or highly singular

. eret post b1     V1

Any idea ???
valérie

________________________________
From

  "Nick Cox" <n.j.cox@durham.ac.uk<mailto:n.j.cox%40durham.ac.uk>>

To

  <statalist@hsphsun2.harvard.edu<mailto:statalist%40hsphsun2.harvard.edu>>

Subject

  st: RE: Problem with "ereturn post" command

Date

  Fri, 5 Feb 2010 12:16:14 -0000

________________________________

The weakest link here that I can identify is the assertion that the

matrices are the same. It's true that they look identical to the number

of decimal places displayed, at least the bits I compared, but that's

not a proof. One thing to explore is thus to calculate the differences

and verify that they really are zero in all elements.



Nick

n.j.cox@durham.ac.uk



Valerie Orozco



I have two matrix, b1 and V1, which respectively correspond to a

coefficient

matrix and variance covariance matrix and want  to use the "ereturn

post'

 command.

I also have matrix Sameb1 and SameV1 (obtained from matrix manipulation

that

I don't show here). They are the same than matrix b1 and V1.

Trying to use "ereturn post" command with Sameb1 and SameV1, Stata tells

me

 I have a problem (Warning:  variance matrix is nonsymmetric or highly

singular) whereas running the command "ereturn post" on b1 and V1 works

well...

See the matrix below...

Does someone have an idea???





. mat list Sameb1



Sameb1[1,8]

         e1vs3       e1vs1      eh1vs1       e2vs3       e2vs2

eh2vs2       e2vs1       e1vs2

r1   .54089558  -.59759218  -.48351881   1.1227012  -1.0151528

-.12922622  -.10754833   .05669663



. mat list b1



b1[1,8]

         e1vs3       e1vs1      eh1vs1       e2vs3       e2vs2

eh2vs2       e2vs1       e1vs2

r1   .54089558  -.59759218  -.48351881   1.1227012  -1.0151528

-.12922622  -.10754833   .05669663



. mat list SameV1



symmetric SameV1[8,8]

             e1vs3       e1vs1      eh1vs1       e2vs3       e2vs2

eh2vs2       e2vs1       e1vs2

 e1vs3   .00126543

 e1vs1   .00399013   .01289885

eh1vs1   .00425701   .01374035   .01463815

 e2vs3   -.0003382  -.00106641  -.00113774   .00009039

 e2vs2   .00140461   .00451378   .00481001   -.0003754   .00158176

eh2vs2   .00113774   .00367227   .00391222  -.00030407   .00128553

.00104559

 e2vs1  -.00106641  -.00344737  -.00367227   .00028501  -.00120636

-.00098146   .00092135

 e1vs2  -.00525556  -.01688898  -.01799736   .00140461  -.00591839

-.00481001   .00451378   .02214454



. mat list V1



symmetric V1[8,8]

             e1vs3       e1vs1      eh1vs1       e2vs3       e2vs2

eh2vs2       e2vs1       e1vs2

 e1vs3   .00126543

 e1vs1   .00399013   .01289885

eh1vs1   .00425701   .01374035   .01463815

 e2vs3   -.0003382  -.00106641  -.00113774   .00009039

 e2vs2   .00140461   .00451378   .00481001   -.0003754   .00158176

eh2vs2   .00113774   .00367227   .00391222  -.00030407   .00128553

.00104559

 e2vs1  -.00106641  -.00344737  -.00367227   .00028501  -.00120636

-.00098146   .00092135

 e1vs2  -.00525556  -.01688898  -.01799736   .00140461  -.00591839

-.00481001   .00451378   .02214454



.

. eret post Sameb1 SameV1

Warning:  variance matrix is nonsymmetric or highly singular



. eret post b1     V1


-------------------------------
Valérie OROZCO
Toulouse School of Economics (INRA-GREMAQ)
21, allée de Brienne
F-31000 Toulouse, France

MA 101
+33 5 61 12 85 91
-------------------------------


*
*   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/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index