performance
performance copied to clipboard
`check_model` and `plot.see_check_*` functions
Maybe it's just me, but the method for check_model
seems really hard to expand / add to currently.
Would it not be easier to have check_model
be a simple wrapper around the various plot.see_check_*
functions, like this:
check_model <- function(model, ...) {
plot(check_normality(model)) +
plot(check_homogeneity(model)) +
plot(check_collinearity(model)) +
plot(check_outliers(model)) +
patchwork::plot_layout(ncol = 2)
}
With the plotting methods for each check_*
living peacefully in see?
(If it is just me, then I need someone to implement check_model.htest
😒)
Yeah, I think is due to that this function has grown over time, and the plot methods for each single functions were partly added later... We could refactor that function, to become easier to maintain.
Yes, it's exactly what Daniel says. We should move these to methods for the specific check functions and also use visualization_recipe().
Sounds like something for a hackathon.......