Keep a given parameter fixed to a predetermined value

[alvarogutyerrez]

Hi Cristian Arteaga,

I was wondering if your program allows you to set a subset of parameters to a given value while optimizing the rest of the parameters.

In particular, I am trying to set the cost coefficients to -1 to immediately read the Value of Time (VoT) from time coefficients.

Is that possible?

Best regards! Álvaro

[arteagac]

Dear @alvarogutyerrez ,

I apologize for my delayed response. I recently implemented a feature that I think facilitates keeping a fixed parameter to a predetermined value. Please see the discussion in the thread #17, which shows how to use this new feature and provides instructions to install the pre-release version that includes this feature. After testing this feature, I will release in the stable version of the package. You are welcome to provide any comments in case you have the chance to test this feature.

[arteagac]

Hello @alvarogutyerrez, I just wanted to let you know that the latest version of xlogit allows you to specify an additive term that helps to model coefficients kept at a fixed value as well as alternative specific fixed values. The additive term can be included in the fit and predict functions as shown below:

model.fit(..., addit= -1*df['cost'], ...)

This additive term is equivalent to a utility specification with a coefficient of -1 kept fixed throughout the estimation for the cost variable (the portion in squared brackets is the addit term):

$$ U_{j} = \beta_0 + \beta_{1j} x_{1j} + ... + [-1\cdot x_{cost j}]$$

Let me know in case you have questions about this functionality.