MNL with random panel effect

[Kevin Dj]

Hello Prof. Bierlaire,

I am trying to build MNL model with a random panel effect. In other words, we want to introduce a random variable (alpha_in) which varies across individuals (n) but not for observation/time (t). In other words, I have the utility function as follows:

U_nit = V_nit + alpha_in + e_nit

I built the models based on the examples provided from the documentation, but I have a suspicion that the Draws used during the simulation for alpha_in might also vary over t. 

My questions are:

1. Is there a settings for the way draws are gathered? (since we might build MXL with the random panel effect as well to capture both heterogeneities (over time and correlation patterns).
2. Does PanelLikelihoodTrajectory() function already adjust the settings for the Draws?

It would be much appreciated if you could help me solve the issue. Thank you so much in advance.

Best,

Kevin Djoenneady Poetera

[Michel Bierlaire]

If you declare the data "panel", the draws are generated per individual.
See an example here: https://biogeme.epfl.ch/sphinx/auto_examples/swissmetro/plot_b12_panel.html#sphx-glr-auto-examples-swissmetro-plot-b12-panel-py

> --
> You received this message because you are subscribed to the Google Groups "Biogeme" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to biogeme+u...@googlegroups.com.
> To view this discussion visit https://groups.google.com/d/msgid/biogeme/fce1e7cf-a2d4-45f2-b319-3e8c08d0e4e2n%40googlegroups.com.

Michel Bierlaire
Transport and Mobility Laboratory
School of Architecture, Civil and Environmental Engineering
EPFL - Ecole Polytechnique Fédérale de Lausanne
http://transp-or.epfl.ch
http://people.epfl.ch/michel.bierlaire

[Kevin Dj]

Hello Prof. Bierlaire,

Thank you for your reply. I have a follow-up question for the MXL with random panel effects. I want to use error components specification for describing correlation patterns, while maintaining the random panel effects. 

When I add the new random variable, it seems that the draws are still generated per individual for both of them.

MNL with panel:

MXL with panel:

Note: other estimation values are identical.

Is there a setting in BIOGEME, where I could set the draws separately for each random variable (i.e., one varies over individual and the other varies over individual and time)?

Best,

Kevin Djoenneady Poetera