Which R-squared should I take?

[yufei xu]

Hi Professor Bierlaire,

I got two R-squared from the estimation report as shown in the attached figure, one is the R-squared for the null model and the other is the R-squared for the initial model. Which one should I use? Or do I need to calculate the R-squared by myself?  

Thanks,

Yufei

[Bierlaire Michel]

Note that it is not a R-squared, like for linear regression. It is just a quantity useful to compare models. It does not mean anything as such. 

My advice is to use AIC and BIC, instead. 

If you really want tp use rho bar square, use the one compared to the null loglikelihood, which is the one usually reported. 

--
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 on the web visit https://groups.google.com/d/msgid/biogeme/478c454f-0849-47ea-ad17-d864ce7c298an%40googlegroups.com.
<Picture1.png>