Pooling models from "CART" imputed data, with zero-inflated negative binomial model

[tingski5]

Hello,

I did just have one other question from mice. If I'm running multiple imputation on a dataset with method "CART," and then I want to use a zero-inflated negative binomial model on the different sets using:

imp <- with(imp1, zeroinfl( ))

It works, but when I go to pool the results, or even look at a summary of it, I get:

Error: No tidy method for objects of class zeroinfl (for summary(imp))
Error: No glance method for objects of class zeroinfl (for summary(pool(imp))

Just wondering if there's a work around to this? With the "with" statement working and me being able to produce different models on the different imputed sets, why couldn't I be able to go further and pool the results?

Thanks for all your help!

[gerkovink]

mice uses broom while pooling since version 3.0.

See #121, #142, #148 and #199 for some more detail.

All the best,

Gerko

[stefvanbuuren]

See https://broom.tidyverse.org/articles/adding-tidiers.html how to create your own glance and tidy methods.

Already available methods:

• broom: https://broom.tidyverse.org/articles/available-methods.html
• broom.mixed: https://github.com/bbolker/broom.mixed