Add "weights" parameter to API for lmer

[sandy1729-swish]

Would it be possible to add access to the weights field in the lmer API?

[ejolly]

Hi @sandy1729-swish can you provide a bit more context for what you mean by the weights field? Model coefficients are accessible via:

model.coef - "population" (top) level estimates
model.fixef - estimates for each group/cluster, i.e. each term after the | in the model formula
model.ranef - deviation estimates for each group, i.e. model.coef - model.fixef

[sandy1729-swish]

Hi Eshin - What I was hoping to access is the the "weights" part of the R API referenced at the top of page 52 in the cran R pdf for lme4: https://cran.r-project.org/web/packages/lme4/lme4.pdf (See the top of page 52.) Does that make more sense? Sandy

…

On Fri, Sep 4, 2020 at 3:23 PM Eshin Jolly ***@***.***> wrote: Hi @sandy1729-swish <https://github.com/sandy1729-swish> can you provide a bit more context for what you mean by the *weights* field? Model coefficients are accessible via: model.coef - "population" (top) level estimates model.fixef - estimates for each group/cluster, i.e. each term after the | in the model formula model.ranef - deviation estimates for each group, i.e. model.coef - model.fixef — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#71 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AN6IRCAMCKVYI3KXLS2HPYLSEFLELANCNFSM4QT6RLFQ> .

[ejolly]

Ah I see thanks for clarifying. At this time I don't have any plans to add that, but can take look if/when I get time. But because it's low on my priority list right now, if it's something you urgently need feel free to open a PR!

[ejolly]

Note: attempted this in #106, but didn't commit a change because it's a bit harder than it looks. Passing a numpy array or list to weights in lmer.lmer doesn't work without some additional conversion for some reason. Will need to take a look later...

[christopherjluke]

Hey Eshin, I was trying to work through a project that needs to be fit in a Python environment so came across your package but the lme4 model I use requires weights as mentioned. Was wondering if you've had taken a look lately on this? I also tried to find a quick and dirty way to pass weights through lmer.lmer but to no avail.

[ejolly]

Hey @christopherjluke unfortunately I just haven't had time. Apologies