Describe the solution you'd like
I would like to be able to plot predictions of quantile regression similarly, like it's currently possible for OLS / GLM. A single quantile would be great, all quantiles at once would be terrific!
It works with only plot_model(model), but doesn't work with plot_model(model, type = "eff", terms = "predictor")
By the way, the emmeans guys (almost, some bugs are still there) made it work with quantile regression and I can use the emmip for plotting. But I prefer your package.
How could we do it?
Should be very similar to OLS with the use of "tau" argument, where tau is the quantile. A list of taus is of couarse a much better choice, a compared to a single tau.
Describe the solution you'd like I would like to be able to plot predictions of quantile regression similarly, like it's currently possible for OLS / GLM. A single quantile would be great, all quantiles at once would be terrific!
It works with only plot_model(model), but doesn't work with plot_model(model, type = "eff", terms = "predictor")
By the way, the emmeans guys (almost, some bugs are still there) made it work with quantile regression and I can use the emmip for plotting. But I prefer your package.
How could we do it? Should be very similar to OLS with the use of "tau" argument, where tau is the quantile. A list of taus is of couarse a much better choice, a compared to a single tau.
Code example
library(ISLR) library(quantreg)
lr <- lm(wage ~ jobclass + age, Wage) qr10 <- rq(wage ~ jobclass + age, Wage, tau = 0.1) qr50 <- rq(wage ~ jobclass + age, Wage, tau = 0.5) qr90 <- rq(wage ~ jobclass + age, Wage, tau = 0.9)
qr_all <- rq(wage ~ jobclass + age, Wage, tau = seq(.05, .95, by = 0.05))
summary(qr_all) %>% plot()