Skip to contents

Fits a generalized linear model (GLM) with quasibinomial family for transition modeling. The quasibinomial family is recommended over binomial as it better handles overdispersion in the data.

Usage

fit_glm(data, ...)

Arguments

data

A data.table containing the result column and predictor columns (prefixed with "id_pred_")

...

Additional arguments (currently ignored, for future extensibility)

Value

A fitted GLM model object, optionally butchered to reduce memory footprint

Details

The function:

  • Uses quasibinomial family to handle overdispersion

  • Automatically detects predictor columns (those starting with "id_pred_")

  • Applies butcher::butcher() if the package is available to reduce model size