hturner / PlackettLuce

PlackettLuce package for Plackett-Luce models in R
https://hturner.github.io/PlackettLuce/
18 stars 5 forks source link

Goodness of fit metric for pltree #52

Closed dbrownf closed 2 years ago

dbrownf commented 3 years ago

Not an issue, but I would appreciate your suggestion on this.

What would be an appropriate goodness of fit metric for a pltree model, especially to asses prediction power with out-of-sample data? Is a pseudo R2 (e.g. McFadden, Maddala, etc.) a suitable option? Or is better to use an accuracy metric comparing predicted vs observed rankings?

Many thanks.