[Limdep Nlogit List] RE: RPL question on specifying starting values
Maksym Polyakov
polyama at auburn.edu
Tue Feb 6 05:25:59 EST 2007
I have a question about specifying starting values for a model like
Brownstone & Train (J. Econometrics, 1999)
mentioned by Iain Pardoe
http://limdep.itls.usyd.edu.au/pipermail/limdep/2007-January/000210.html :
> NLOGIT ; Lhs = ch
> ; Choices = A1,A2,B1,B2,C1,C2
> ; Rhs = price,rang1,acc1,spee1,poll,size1,big,lugg,cost1,stav,
>
> suv,spo,wag,tru,van,ev,noev,comev,colev,cng,nocng,meth,colmet
> ; RPL
> ; Fcn = noev(N,*,0),nocng(N,*,0),size1(N),lugg(N)
> ; Halton
> ; Pts = 100 $
>
There are means of 23 coefficients (2 of which are restricted), and
variances of 4 coefficients. LIMDEP accepts 27 parameters for this
model. When I tried to provide parameters in the same order they appear
in output, tracing using output=3 had shown that LIMDEP ignored first
four parameters and duplicated last two. What is the correct way to
specify starting values?
Code:
NLOGIT ; Lhs = ch
; Choices = A1,A2,B1,B2,C1,C2
; Rhs = price,rang1,acc1,spee1,poll,size1,big,lugg,cost1,stav,
suv,spo,wag,tru,van,ev,noev,comev,colev,cng,nocng,meth,colmet
; RPL
; Fcn = noev(N,*,0),nocng(N,*,0),size1(N),lugg(N)
; Halton
; Pts = 100
;output = 3
; start= .000000, .000000, 1.39293693, 1.81442959, -.26872126,
53035615, -1.11834103,
.30670273, -.61532464, .23622439, -1.27993747, .62350702,
89638305, .70185209,
-1.50438859, -1.09073195, -.81790292, -1.36276549, .40458394,
85239901, .65000510,
-.42018760, .31061471, 2.95182904, 1.23502550, 7.12455487,
6.28112112 $
Fragment of output:
Start values: -.26872D+00 .53036D+00 -.11183D+01 .30670D+00
-.61532D+00
.23622D+00 -.12799D+01 .62351D+00 .89638D+00 .70185D+00
-.15044D+01
-.10907D+01 -.81790D+00 -.13628D+01 .40458D+00 .85240D+00
65001D+00
-.42019D+00 .31061D+00 .29518D+01 .12350D+01 .71246D+01
62811D+01
.71246D+01 .62811D+01
--
Maksym Polyakov
Research Fellow
School of Forestry and Wildlife Sciences
Auburn University, AL 36849-5418
phone: (334) 844 8061
fax: (334) 844 1084
More information about the Limdep
mailing list