Symbolize a base R lm() fit (Gaussian, v0.10 first slice)

Builds a symbolized_model from a base R lm object. v0.10 covers the Gaussian conditional submodel only (the only family lm() fits). The residual standard deviation (summary(fit)$sigma) is exposed as the sigma variance component but is not modelled – lm() doesn't support distributional regression on sigma.

Usage

# S3 method for class 'lm'
symbolize(fit, symbols = NULL, units = NULL, context = NULL, ...)

Arguments

fit: A fitted statistical model object.
symbols: Optional named character vector mapping variable names to user-supplied LaTeX symbols, e.g. c(body_mass = "W_i", temperature = "T_i").
units: Optional named character vector mapping variable names to units, e.g. c(body_mass = "g", temperature = "C").
context: Optional short character description of the model, e.g. "avian body-size location-scale model".
...: Reserved for method-specific extra arguments.

Value

A symbolized_model object.

Confidence intervals

CIs come from stats::confint(fit). The method is "Wald-t" – the usual frequentist t-based CI you'd see in summary(fit). ci_method is set to "wald" for consistency with the other frequentist extractors.

References

Chambers, J. M. (1992). Linear models. Chapter 4 of Statistical Models in S (Chambers, J. M., & Hastie, T. J., eds.). Wadsworth & Brooks/Cole.