Re: Estimating Destination choice model with one size variable

[Bierlaire Michel]

Dear Hend,

Can you please post your questions on the users’ group? 

I did not see the attachment. If the init model is equal to zero, it is likely that the data are not read properly, probably because it is not in the right format.

If you are using biogeme 2 (called python biogeme), it is not maintained anymore. You should move to biogeme 3 (called pandas biogeme) which is a python package. 

In that case, the data are managed by the pandas library which is powerful and flexible.

On 1 Jun 2022, at 12:26, Hend Manasra <he...@enmotion.co.il> wrote:

Dear  Prof. M. Bierlaire,

I am trying to estimate a destination choice model for a metropolitan area according to a survey.

The utility function is like:   U = a1x1+ a2x2... + log(a3x3), when x3 is the size variable.

When I implemented the model in python Biogeme and tried to run the estimation using results = biogeme.estimate() the model run but I got nulls for all of the coefficients I want to estimate (a1, a2, a3) and the Rho-square for the init. model was equal to zero (attached the result file).

When I estimate the model with two size variables when one of them is fixed to 1, the model runs successfully with logical results.

 

My question is if I should estimate a model with at least one fixed value? and how can I estimate a model with just one size variable?

Many thanks

Hend

--