support for `quantreg::rq` objects

[fawda123]

Hi there, great package! Just a suggestion to add support for quantreg::rq objects.

library(equatiomatic)
library(quantreg)

# Fit a simple model
mod1 <- rq(mpg ~ cyl + disp, data = mtcars)

# Give the results to extract_eq
extract_eq(mod1)

# Error in UseMethod("extract_lhs", model) : 
#  no applicable method for 'extract_lhs' applied to an object of class "rq"

[datalorax]

Hi, thanks for the suggestion! This is a great idea but, honestly, we're probably a ways away from being able to implement it (and I recognize this issue was opened nearly 11 months ago with no response). I'm hoping to build out a framework to make extensions like this easier, and then to build a vignette on how to implement new equations. That way, others in the community like yourself could help contribute to the package more directly. I'll tag this issue if/when I'm able to do that.

[fawda123]

No worries, feel free to tag me with any updates!

[datalorax]

Hi @fawda123, could you provide an example of what the equation from the model you showed above should look like? I'm not super familiar with quantile regression.

[fawda123]

@datalorax the equation is the same as for OLS, i.e., response ~ intercept plus slope parameters.

library(quantreg)
mod1 <- rq(mpg ~ cyl + disp, data = mtcars)
summary(mod1)

# Call: rq(formula = mpg ~ cyl + disp, data = mtcars)
# 
# tau: [1] 0.5
# 
# Coefficients:
#   coefficients lower bd upper bd
# (Intercept) 33.32282     27.32925 38.72945
# cyl         -1.58246     -2.66538 -0.55182
# disp        -0.01768     -0.03105 -0.00331

[datalorax]

Ah... okay. Then yes! This should be pretty easy to implement.