How to estimate a Generalised Nested Logit model by GAUSS 17?

Hello everyone. In my PhD, I try to introduce the joint choice among departure time, destination and mode. I purpose some NL structures and I already estimated 2-Levels, 3-Levels and even 4-Level NL models gy using NLogit Software. I also estimated 2-Levels NL-OGEV model in which ordered nature of adjacent destinations are introduced. However, my main point of interest is the estimation of a GNL model with general nest and number of nests according to the potential correlation of departure time, destination and mode. So far, I could estimate it by Nlogit but with many assumptions and restrictions related to inclusive value parameters. I have purchased GAUSS 17 but I countered difficulties in learning its coding the model. Could any one help me to code such a model with GAUSS 17 or any other available software with complete rights of mention his/her participation in any papers of the thesis? I am ready to share required data and the estimated models and any other required information.

3 Answers



0



accepted

I would suggest taking a look at several papers by Professor Habib at the University of Toronto:

https://www.researchgate.net/profile/Khandker_Nurul_Habib2/publications

 

He estimates several similar models and they are generally estimated in GAUSS. He is my supervisor, so I could connect you if his models prove helpful.



0



Sorry, I should mention that the main steps are:

  1. Setting up the data in GAUSS (good documentation online)
  2. Specifying the log likelihood function (GAUSS has a good built-in function). I would suggest looking through several papers with similar likelihood functions.

I have a few models on my Github:

https://github.com/jfhawkin

and Chandra Bhat posts his estimation code on his website:

http://www.ce.utexas.edu/prof/bhat/FULL_CODES.htm



0



Dear Hawkin,

I appreciate your kindly interest. I have met with professor Hbib through the last TRB. He looks a gentleman and so kind. I will be so glad if you can provide connection between me and him to achieve my goal. My e-mail: [email protected]. Thank you for your help.

Your Answer

3 Answers

0
accepted

I would suggest taking a look at several papers by Professor Habib at the University of Toronto:

https://www.researchgate.net/profile/Khandker_Nurul_Habib2/publications

 

He estimates several similar models and they are generally estimated in GAUSS. He is my supervisor, so I could connect you if his models prove helpful.

0

Sorry, I should mention that the main steps are:

  1. Setting up the data in GAUSS (good documentation online)
  2. Specifying the log likelihood function (GAUSS has a good built-in function). I would suggest looking through several papers with similar likelihood functions.

I have a few models on my Github:

https://github.com/jfhawkin

and Chandra Bhat posts his estimation code on his website:

http://www.ce.utexas.edu/prof/bhat/FULL_CODES.htm

0

Dear Hawkin,

I appreciate your kindly interest. I have met with professor Hbib through the last TRB. He looks a gentleman and so kind. I will be so glad if you can provide connection between me and him to achieve my goal. My e-mail: [email protected]. Thank you for your help.


You must login to post answers.

Have a Specific Question?

Get a real answer from a real person

Need Support?

Get help from our friendly experts.