search

biocore / songbird

Vanilla regression methods for microbiome differential abundance analysis
BSD 3-Clause "New" or "Revised" License
58 stars 25 forks source link

Multinomial regression continuous variable scaling #135

Closed ibrila closed 3 years ago

ibrila commented 4 years ago

Hi,

I'm planning to use Songbird multinomial model with multiple variables that are on quite different scales (e.g. 0.000092 and 4.465734). I've done scaling of the same variables when running a linear mixed model, and was wondering if this type of model needs variable rescaling prior to running the model, and is there a recommended method for this?

mortonjt commented 4 years ago

Yes, standardizing your covariates is generally a good practice. See

https://statmodeling.stat.columbia.edu/2009/07/11/when_to_standar/

On Wed, Aug 12, 2020, 2:20 PM Ilze notifications@github.com wrote:

Hi,

I'm planning to use Songbird multinomial model with multiple variables that are on quite different scales (e.g. 0.000092 and 4.465734). I've done scaling of the same variables when running a linear mixed model, and was wondering if this type of model needs variable rescaling prior to running the model, and is there a recommended method for this?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/biocore/songbird/issues/135, or unsubscribe https://github.com/notifications/unsubscribe-auth/AA75VXJSUOBMRXJ6XPYL77TSAL2RJANCNFSM4P5HSROA .

ibrila commented 4 years ago

Thanks for the great resource and help!

mortonjt commented 3 years ago

It looks like this question has been answered, feel free to reopen if this is not the case.