[Limdep Nlogit List] Variable choice sets

Thomas C. Eagle teagle at tceagle.com
Sat Jun 19 06:42:43 EST 2010


Use namelists, 

Or switch to Biogeme.

Tom

-----Original Message-----
From: limdep-bounces at limdep.itls.usyd.edu.au
[mailto:limdep-bounces at limdep.itls.usyd.edu.au] On Behalf Of Yijing Lu
Sent: Friday, June 18, 2010 12:46 PM
To: Limdep and Nlogit Mailing List
Subject: Re: [Limdep Nlogit List] Variable choice sets

I think it's kind of impossible to estimate the MNL by specifying all the
utility functions, since there are 11 attributes varying across the
alternatives and 20 invariant variables for each alternative and the
universal choice sets consist of more than 50 alternatives. I tried the
method however, got the error: The model is too complex - too many parts.
Table overflow. or  Error  2: Command is more than 2500 characters.

YL

2010/6/18 Thomas C. Eagle <teagle at tceagle.com>

> Below is an example of a MNL model with choice sets varying in composition
> with the utility function of each alternative is fully specified (alts
a122
> thru a135 could be absence in any specific choice set).  Using this
> formulation the choice sets can consist of any subset or all alternatives.
> I think you need a set up like it.
>
> Tom Eagle
>
> ?  MNL model underlying current (Jan 31, 2008) nested logit
>
> NLOGIT
> ;Frequencies
> ;Lhs=VOLUME, NALTS, ALT
> ;Choices= a122, a123, a124, a125, a126, a127, a128, a129, a130, a131,
>    a132, a133, a134, a135
> ;Maxit = 300
> ;Model:
>    U(a122) =          bpr122 * lnpr + b122_1 * dma_m1 + b122_2 * dma_m2
>    + b122_3_7 * dma_m3 + b122_3_7 * dma_m5 + b122_3_7 * dma_m6
>    + b122_3_7 * dma_m7
>    /
>    U(a123) = asc123 + bpr123 * lnpr + b123_1 * dma_m1 + b123_2 * dma_m2
>    + b123_3 * dma_m3 + b123_5 * dma_m5 + b123_6 * dma_m6 + b123_7 * dma_m7
>    /
>    U(a124) = asc124 + bpr124 * lnpr + b124_2 * dma_m2 + b124_3 * dma_m3
>    + b124_46 * dma_m4 + b124_5 * dma_m5 + b124_46 * dma_m6
>    + b124_7 * dma_m7
>    /
>    U(a125) = asc125 + bpr125 * lnpr + b125_2 * dma_m2
>    + b125_46 * dma_m4 + b125_5 * dma_m5 + b125_46 * dma_m6
>    + b125_7 * dma_m7
>    /
>    U(a126) = asc126 + bpr126 * lnpr + bf126 * FeaD + b126_15 * dma_m1
>    + b126_2 * dma_m2 + b126_4 * dma_m4
>    + b126_15 * dma_m5 + b126_6 * dma_m6 + b126_7 * dma_m7
>    /
>    U(a127) = asc127 + bpr127 * lnpr + bf127 * FeaD + b127_1 * dma_m1
>    + b127_24 * dma_m2 + b127_3 * dma_m3 + b127_24 * dma_m4
>    + b127_7 * dma_m7
>    /
>    U(a128) = asc128 + bpr128 * lnpr + b128_4 * dma_m4 + b128_5 * dma_m5
>    + b128_6 * dma_m6
>    /
>    U(a129) =          bpr129 * lnpr
>    /
>    U(a130) = asc130 + bpr130 * lnpr + b130_1 * dma_m1 + b130_2 * dma_m2
>    + b130_3 * dma_m3 + b130_4 * dma_m4 + b130_5 * dma_m5
>    + b130_6 * dma_m6 + b130_7 * dma_m7
>    /
>    U(a131) = asc131 + bpr131 * lnpr + b131_24 * dma_m2
>    + b131_35 * dma_m3 + b131_24 * dma_m4 + b131_35 * dma_m5
>    + b131_7 * dma_m7
>    /
>    U(a132) = asc132 + bpr132 * lnpr+ b132_1 * dma_m1 + b132_2 * dma_m2
>    + b132_35 * dma_m3 + b132_46 * dma_m4 + b132_35 * dma_m5
>    + b132_46 * dma_m6 + b132_7 * dma_m7
>    /
>    U(a133) = asc133 + bpr133 * lnpr + b133_1 * dma_m1
>    + b133_2 * dma_m2 + b133_3 * dma_m3 + b133_6 * dma_m6
>    + b133_7 * dma_m7  /
>    U(a134) = asc134 + bpr134 * lnpr + bf134 * FeaD + b134_17 * dma_m1
>    + b134_2 * dma_m2 + b134_3 * dma_m3 + b134_4 * dma_m4
>    + b134_5 * dma_m5 + b134_6 * dma_m6 + b134_17 * dma_m7
>    /
>    U(a135) = asc135 + bpr135 * lnpr + bf135 * FeaD  + b135_2 * dma_m2
>
> $
>
>
> _______________________________________________
> Limdep site list
> Limdep at limdep.itls.usyd.edu.au
> http://limdep.itls.usyd.edu.au
>
_______________________________________________
Limdep site list
Limdep at limdep.itls.usyd.edu.au
http://limdep.itls.usyd.edu.au



More information about the Limdep mailing list